Showing preview only (284K chars total). Download the full file or copy to clipboard to get everything.
Repository: GewelsJI/VPS
Branch: main
Commit: 4b2d87e8265a
Files: 68
Total size: 263.4 KB
Directory structure:
gitextract_qn90sj5e/
├── .gitignore
├── LICENSE
├── README.md
├── docs/
│ ├── AWESOME_VPS.md
│ ├── DATA_DESCRIPTION.md
│ ├── DATA_PREPARATION.md
│ ├── INFO_NEGATIVE_CASES.md
│ ├── INFO_POSITIVE_CASES.md
│ └── RELEASE_NOTES.md
├── eval/
│ ├── README.md
│ ├── eval-result/
│ │ ├── 2015-MICCAI-UNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2018-TMI-UNet++/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2019-TPAMI-COSNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2020-AAAI-PCSA/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2020-MICCAI-23DCNN/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2020-MICCAI-ACSNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2020-MICCAI-PraNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2020-TIP-MATNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2021-ICCV-DCFNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2021-ICCV-FSNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2021-MICCAI-PNSNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2021-MICCAI-SANet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2021-NIPS-AMD/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ └── 2022-TMI-PNSPlus/
│ │ ├── TestEasyDataset_eval.txt
│ │ └── TestHardDataset_eval.txt
│ ├── eval.sh
│ ├── metrics.py
│ └── vps_evaluator.py
├── lib/
│ ├── __init__.py
│ ├── dataloader/
│ │ ├── __init__.py
│ │ ├── dataloader.py
│ │ ├── preprocess.py
│ │ └── statistics.pth
│ ├── module/
│ │ ├── LightRFB.py
│ │ ├── PNS/
│ │ │ ├── PNS_Module/
│ │ │ │ ├── CMakeLists.txt
│ │ │ │ ├── reference.cpp
│ │ │ │ ├── reference.h
│ │ │ │ ├── sa.cu
│ │ │ │ ├── sa_ext.cpp
│ │ │ │ ├── timer.h
│ │ │ │ └── utils.h
│ │ │ └── setup.py
│ │ ├── PNSPlusModule.py
│ │ ├── PNSPlusNetwork.py
│ │ ├── Res2Net_v1b.py
│ │ └── __init__.py
│ └── utils/
│ ├── __init__.py
│ └── utils.py
├── scripts/
│ ├── config.py
│ ├── eval_eff.py
│ ├── my_test.py
│ └── my_train.py
├── snapshot/
│ └── .placeholder
└── utils/
├── reorganize.py
└── reorganize.sh
================================================
FILE CONTENTS
================================================
================================================
FILE: .gitignore
================================================
# For projects using Nanoc (http://nanoc.ws/)
# Default location for output (needs to match output_dir's value found in nanoc.yaml)
output/
# Temporary file directory
tmp/nanoc/
# Crash Log
crash.log
data/
res/
eval/loggings/
eval/__pycache__/
backup/
./TODOLIST.md
================================================
FILE: LICENSE
================================================
Apache License
Version 2.0, January 2004
http://www.apache.org/licenses/
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
1. Definitions.
"License" shall mean the terms and conditions for use, reproduction,
and distribution as defined by Sections 1 through 9 of this document.
"Licensor" shall mean the copyright owner or entity authorized by
the copyright owner that is granting the License.
"Legal Entity" shall mean the union of the acting entity and all
other entities that control, are controlled by, or are under common
control with that entity. For the purposes of this definition,
"control" means (i) the power, direct or indirect, to cause the
direction or management of such entity, whether by contract or
otherwise, or (ii) ownership of fifty percent (50%) or more of the
outstanding shares, or (iii) beneficial ownership of such entity.
"You" (or "Your") shall mean an individual or Legal Entity
exercising permissions granted by this License.
"Source" form shall mean the preferred form for making modifications,
including but not limited to software source code, documentation
source, and configuration files.
"Object" form shall mean any form resulting from mechanical
transformation or translation of a Source form, including but
not limited to compiled object code, generated documentation,
and conversions to other media types.
"Work" shall mean the work of authorship, whether in Source or
Object form, made available under the License, as indicated by a
copyright notice that is included in or attached to the work
(an example is provided in the Appendix below).
"Derivative Works" shall mean any work, whether in Source or Object
form, that is based on (or derived from) the Work and for which the
editorial revisions, annotations, elaborations, or other modifications
represent, as a whole, an original work of authorship. For the purposes
of this License, Derivative Works shall not include works that remain
separable from, or merely link (or bind by name) to the interfaces of,
the Work and Derivative Works thereof.
"Contribution" shall mean any work of authorship, including
the original version of the Work and any modifications or additions
to that Work or Derivative Works thereof, that is intentionally
submitted to Licensor for inclusion in the Work by the copyright owner
or by an individual or Legal Entity authorized to submit on behalf of
the copyright owner. For the purposes of this definition, "submitted"
means any form of electronic, verbal, or written communication sent
to the Licensor or its representatives, including but not limited to
communication on electronic mailing lists, source code control systems,
and issue tracking systems that are managed by, or on behalf of, the
Licensor for the purpose of discussing and improving the Work, but
excluding communication that is conspicuously marked or otherwise
designated in writing by the copyright owner as "Not a Contribution."
"Contributor" shall mean Licensor and any individual or Legal Entity
on behalf of whom a Contribution has been received by Licensor and
subsequently incorporated within the Work.
2. Grant of Copyright License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
copyright license to reproduce, prepare Derivative Works of,
publicly display, publicly perform, sublicense, and distribute the
Work and such Derivative Works in Source or Object form.
3. Grant of Patent License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
(except as stated in this section) patent license to make, have made,
use, offer to sell, sell, import, and otherwise transfer the Work,
where such license applies only to those patent claims licensable
by such Contributor that are necessarily infringed by their
Contribution(s) alone or by combination of their Contribution(s)
with the Work to which such Contribution(s) was submitted. If You
institute patent litigation against any entity (including a
cross-claim or counterclaim in a lawsuit) alleging that the Work
or a Contribution incorporated within the Work constitutes direct
or contributory patent infringement, then any patent licenses
granted to You under this License for that Work shall terminate
as of the date such litigation is filed.
4. Redistribution. You may reproduce and distribute copies of the
Work or Derivative Works thereof in any medium, with or without
modifications, and in Source or Object form, provided that You
meet the following conditions:
(a) You must give any other recipients of the Work or
Derivative Works a copy of this License; and
(b) You must cause any modified files to carry prominent notices
stating that You changed the files; and
(c) You must retain, in the Source form of any Derivative Works
that You distribute, all copyright, patent, trademark, and
attribution notices from the Source form of the Work,
excluding those notices that do not pertain to any part of
the Derivative Works; and
(d) If the Work includes a "NOTICE" text file as part of its
distribution, then any Derivative Works that You distribute must
include a readable copy of the attribution notices contained
within such NOTICE file, excluding those notices that do not
pertain to any part of the Derivative Works, in at least one
of the following places: within a NOTICE text file distributed
as part of the Derivative Works; within the Source form or
documentation, if provided along with the Derivative Works; or,
within a display generated by the Derivative Works, if and
wherever such third-party notices normally appear. The contents
of the NOTICE file are for informational purposes only and
do not modify the License. You may add Your own attribution
notices within Derivative Works that You distribute, alongside
or as an addendum to the NOTICE text from the Work, provided
that such additional attribution notices cannot be construed
as modifying the License.
You may add Your own copyright statement to Your modifications and
may provide additional or different license terms and conditions
for use, reproduction, or distribution of Your modifications, or
for any such Derivative Works as a whole, provided Your use,
reproduction, and distribution of the Work otherwise complies with
the conditions stated in this License.
5. Submission of Contributions. Unless You explicitly state otherwise,
any Contribution intentionally submitted for inclusion in the Work
by You to the Licensor shall be under the terms and conditions of
this License, without any additional terms or conditions.
Notwithstanding the above, nothing herein shall supersede or modify
the terms of any separate license agreement you may have executed
with Licensor regarding such Contributions.
6. Trademarks. This License does not grant permission to use the trade
names, trademarks, service marks, or product names of the Licensor,
except as required for reasonable and customary use in describing the
origin of the Work and reproducing the content of the NOTICE file.
7. Disclaimer of Warranty. Unless required by applicable law or
agreed to in writing, Licensor provides the Work (and each
Contributor provides its Contributions) on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied, including, without limitation, any warranties or conditions
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
PARTICULAR PURPOSE. You are solely responsible for determining the
appropriateness of using or redistributing the Work and assume any
risks associated with Your exercise of permissions under this License.
8. Limitation of Liability. In no event and under no legal theory,
whether in tort (including negligence), contract, or otherwise,
unless required by applicable law (such as deliberate and grossly
negligent acts) or agreed to in writing, shall any Contributor be
liable to You for damages, including any direct, indirect, special,
incidental, or consequential damages of any character arising as a
result of this License or out of the use or inability to use the
Work (including but not limited to damages for loss of goodwill,
work stoppage, computer failure or malfunction, or any and all
other commercial damages or losses), even if such Contributor
has been advised of the possibility of such damages.
9. Accepting Warranty or Additional Liability. While redistributing
the Work or Derivative Works thereof, You may choose to offer,
and charge a fee for, acceptance of support, warranty, indemnity,
or other liability obligations and/or rights consistent with this
License. However, in accepting such obligations, You may act only
on Your own behalf and on Your sole responsibility, not on behalf
of any other Contributor, and only if You agree to indemnify,
defend, and hold each Contributor harmless for any liability
incurred by, or claims asserted against, such Contributor by reason
of your accepting any such warranty or additional liability.
END OF TERMS AND CONDITIONS
APPENDIX: How to apply the Apache License to your work.
To apply the Apache License to your work, attach the following
boilerplate notice, with the fields enclosed by brackets "[]"
replaced with your own identifying information. (Don't include
the brackets!) The text should be enclosed in the appropriate
comment syntax for the file format. We also recommend that a
file or class name and description of purpose be included on the
same "printed page" as the copyright notice for easier
identification within third-party archives.
Copyright [yyyy] [name of copyright owner]
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
================================================
FILE: README.md
================================================
# <p align=center>`Video Polyp Segmentation: A Deep Learning Perspective (MIR 2022)`</p><!-- omit in toc -->
[](https://opensource.org/licenses/MIT)
[](https://arxiv.org/pdf/2203.14291v3.pdf)
[](https://gitter.im/video-polyp-segmentation/community?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge)
[](https://paperswithcode.com/sota/video-polyp-segmentation-on-sun-seg-easy?p=video-polyp-segmentation-a-deep-learning)
[](https://paperswithcode.com/sota/video-polyp-segmentation-on-sun-seg-hard?p=video-polyp-segmentation-a-deep-learning)
---
> [!note]
> The original SUN Database website (http://amed8k.sundatabase.org/) is no longer maintained and is sometimes inaccessible, we provide a backup website for the SUN Database here: [website backup](https://github.com/GewelsJI/VPS/blob/main/docs/%5Boriginal%20backup%5D%20SUN%20Colonoscopy%20Video%20Database.pdf)
>
> Regarding recent requests for a large volume of data, we suggest you emailing me at gepengai.ji@gmail.com for alternative access options.
>
> 🔴 We're currently pushing intelligent colonoscopy (refer to our [research gallery](https://github.com/ai4colonoscopy) page) into the multimodal era. Recommen you reading two new research works from our team:
> - :boom: The pioneering multimodal analysis solution for colonoscopy: **ColonINST** & **ColonGPT** ([paper](https://arxiv.org/abs/2410.17241) & [project page](https://github.com/ai4colonoscopy/IntelliScope))
> - :boom: The current largest multimodal dataset **ColonVQA (1.1+ million entries)** in colonoscopy. We also introduce the first reasoning-centric dataset **ColonReason**, along with an R1-styled model, **ColonR1**, tailored for colonoscopy tasks. ([paper](https://arxiv.org/abs/2512.03667) & [project page](https://github.com/ai4colonoscopy/Colon-X))
---
<p align="center">
<img src="./assets/background-min.gif"/> <br />
</p>
- **Title:** Video Polyp Segmentation: A Deep Learning Perspective (accepted by Machine Intelligence Research, please see [arXiv version](https://arxiv.org/pdf/2203.14291v3.pdf) & [Spriner version](https://link.springer.com/article/10.1007/s11633-022-1371-y))
- **Authors:** [Ge-Peng Ji](https://scholar.google.com/citations?view_op=list_works&hl=en&hl=en&user=oaxKYKUAAAAJ)^, [Guobao Xiao](https://guobaoxiao.github.io)^, [Yu-Cheng Chou](https://sites.google.com/view/yu-cheng-chou)^, [Deng-Ping Fan](https://dengpingfan.github.io/)*, [Kai Zhao](https://kaizhao.net/), [Geng Chen](https://scholar.google.com/citations?user=sJGCnjsAAAAJ&hl=en), and [Luc Van Gool](https://scholar.google.com/citations?user=TwMib_QAAAAJ&hl=en).
- **Contact:** We invite all to contribute to making it more accessible and useful. If you have any questions, please feel free to drop us an e-mail (gepengai.ji@gmail.com & dengpfan@gmail.com) or directly report the issue or push a PR. Your star is our motivation, let's enjoy it!
- Welcome any discussions on video polyp segmentation at [Gitter room](https://gitter.im/video-polyp-segmentation/community?utm_source=share-link&utm_medium=link&utm_campaign=share-link) or join our [WeChat group](https://github.com/GewelsJI/VPS/blob/main/assets/wechat_group_qr_code_20240816.JPG).
- The following is a video to quickly access the core points of our work.
https://github.com/GewelsJI/VPS/assets/38354957/9bea01ae-9582-494f-8bf6-f83307eebc08
# Contents<!-- omit in toc -->
- [1. Features](#1-features)
- [2. News](#2-news)
- [3. VPS Dataset](#3-vps-dataset)
- [4. VPS Baseline](#4-vps-baseline)
- [5. VPS Benchmark](#5-vps-benchmark)
- [6. Tracking Trends](#6-tracking-trends)
- [7. Citations](#7-citations)
- [8. FAQ](#8-faq)
- [9. License](#9-license)
- [10. Acknowledgments](#10-acknowledgments)
# 1. Features
In the deep learning era, we present the first comprehensive video polyp segmentation (VPS) study. Over the years, developments on VPS have not moved forward with ease since large-scale fine-grained segmentation masks are still not made publicly available. To tackle this issue, we first introduce a long-awaited high-quality per-frame annotated VPS dataset. There are four features of our work:
- **VPS Dataset:** We recognize the importance of annotated medical data for substantial progress in research on medical AI systems’ development. Thus, our SUN-SEG dataset is open access, a non-profit database of the high-quality, large-scale, densely annotated dataset for facilitating colonoscopy diagnosis, localization, and derivative tasks. Our vision aims to provide data and knowledge to aid and educate clinicians, and also for the development of automated medical decision support systems.
- **VPS Baseline:** We propose a simple but efficient baseline, which outperforms the 13 cutting-edge polyp segmentation approaches and runs in super real-time (170fps). We hope such a baseline could attract more researchers to join our community and inspire them to develop more interesting solutions.
- **VPS Benchmark:** For a fair comparison, we build an online leaderboard to keep up with the new progress of the VPS community. Besides, we provide an out-of-the-box evaluation toolbox for the VPS task.
- **Tracking Trends:** We elaborately collect a paper reading list ( :boom: [Awesome Paper List](https://github.com/GewelsJI/VPS/blob/main/docs/AWESOME_VPS.md) :boom: ) to continuously track the latest updates in this rapidly advancing field.
# 2. News
- *[Jul/15/2024]* Thanks to [@Yuli Zhou](https://github.com/zhoustan) for raising up the frame sorting issues in the inference and evaluation code (check out the pull request here: https://github.com/GewelsJI/VPS/pull/48). It turns out it only slightly impacted our final evaluation performance (check out the evidence [here](https://github.com/GewelsJI/VPS/blob/main/docs/RELEASE_NOTES.md)).
- *[Oct/26/2023]* The video-level attributes have released at [Google Drive](https://docs.google.com/spreadsheets/d/1J33EvrEcZp5CMWtKN_4VNdhp4EQsjMSf/edit?usp=sharing&ouid=117958307137184272405&rtpof=true&sd=true).
- *[Jan/30/2023]* We update the bounding box annotation with COCO-like format, ie, `[x,y,width,height]` where x and y are the upper-left coordinates of the bounding box. Please download the latest compressed file at [here](https://github.com/GewelsJI/VPS/blob/main/docs/DATA_PREPARATION.md#step-1-request-and-download). Thanks for Yingling Lu for pointing out this issue.
- *[August/24/2022]* :boom: Our paper has been accepted by [Machine Intelligence Research (MIR)](https://www.springer.com/journal/11633) journal.
- *[July/03/2022]* :boom: We update a new version of SUN-SEG with more fine-grained data splits, including seen/unseen scenarios. For more details refer to our technical report. Also, the new PaperWithCode page refers to [SUN-SEG-Easy](https://paperswithcode.com/dataset/sun-seg-easy) & [SUN-SEG-Hard](https://paperswithcode.com/dataset/sun-seg-hard).
- *[May/11/2022]* Release rejected labels: [SUN-SEG-Rejected-Labels (Google Drive, 120.7MB)](https://drive.google.com/file/d/1OtK2PR6gKQv56dIFjw0rXadcgGonf93S/view?usp=sharing). For more details see [here](https://github.com/GewelsJI/VPS/blob/main/docs/DATA_DESCRIPTION.md#rejected-labels).
- *[March/27/2022]* Release pretrained checkpoints and whole benchamrks results.
- *[March/18/2022]* Upload the whole training/testing code for our enhanced model PNS+.
- *[March/15/2022]* Release the evaluation toolbox for the VPS task. Add a [Awesome_Video_Polyp_Segmentation.md](https://github.com/GewelsJI/VPS/blob/main/docs/AWESOME_VPS.md) for tracking latest trends of this community.
- *[March/14/2022]* Create the project page.
# 3. VPS Dataset
<p align="center">
<img src="./assets/Pathological-min.gif"/> <br />
<em>
Figure 1: Annotation of SUN-SEG dataset. The object-level segmentation masks in the SUN-SEG dataset of different pathological categories, which is densely annotated with experienced annotators and verified by colonoscopy-related researchers to ensure the quality of the proposed dataset.
</em>
</p>
Notably, based on some necessary privacy-preserving considerations from the SUN dataset, we could not directly share the download link of the video dataset with you without authorization. And please inform us of your institution and the purpose of using SUN-SEG in the email. Thank you for your understanding!
- How to get access to our SUN-SEG dataset? Please refer to [`DATA_PREPARATION`](https://github.com/GewelsJI/VPS/blob/main/docs/DATA_PREPARATION.md).
- If you wanna know more descriptions about our SUN-SEG dataset. Please refer to our [`DATA_DESCRIPTION.md`](https://github.com/GewelsJI/VPS/blob/main/docs/DATA_DESCRIPTION.md).
# 4. VPS Baseline
> This work is the extension version of our conference paper (Progressively Normalized Self-Attention Network for Video Polyp Segmentation) accepted at MICCAI-2021. More details could refer to [arXiv](https://arxiv.org/abs/2105.08468) and [Github Link](https://github.com/GewelsJI/PNS-Net)
<p align="center">
<img src="./assets/PNSPlus-Framework.png"/> <br />
<em>
Figure 2: The pipeline of the proposed (a) PNS+ network, which is based on (b) the normalized self-attention (NS) block.
</em>
</p>
There are three simple-to-use steps to access our project code (PNS+):
- Prerequisites of environment:
```bash
conda create -n PNS+ python=3.6
conda activate PNS+
conda install pytorch=1.1.0 torchvision -c pytorch
pip install tensorboardX tqdm Pillow==6.2.2
pip install git+https://github.com/pytorch/tnt.git@master
```
- Compiling the project:
```bash
cd ./lib/module/PNS
python setup.py build develop
```
- Training:
```bash
python ./scripts/my_train.py
```
- Testing:
Downloading pre-trained weights and move it into `snapshot/PNSPlus/epoch_15/PNSPlus.pth`,
which can be found in this download link: [Google Drive, 102.9MB](https://drive.google.com/file/d/1YCC9AiSr3yMhPXBEnM2bgjoJ7fodpLkK/view?usp=sharing) / [Baidu Drive](https://pan.baidu.com/s/1mnd9GD2BiWFzsibv7WiwAA) (Password: g7sa, Size: 108MB).
```bash
python ./scripts/my_test.py
```
# 5. VPS Benchmark
We provide an out-of-the-box evaluation toolbox for the VPS task, which is written in Python style. You can just run it to generate the evaluation results on your custom approach. Or you can directly download the complete VPS benchmark including the prediction map of each competitor at the download link: [Google Drive, 5.45GB](https://drive.google.com/file/d/1Liva1oR1-1ihWaTNNM5WDKZVETAF587M/view?usp=sharing) / [Baidu Drive](https://pan.baidu.com/s/1Qu-0l9w0ja92nzrlWMRhoQ) (Password: 2t1l, Size: 5.45G).
- More instructions about **Evaluation Toolbox** refer to [`PageLink`](https://github.com/GewelsJI/VPS/tree/main/eval).
We also built an online leaderboard to keep up with the new progress of other competitors. We believe this is a fun way to learn about new research directions and stay in tune with our VPS community.
- Online leaderboard is publicly avaliable at PaperWithCode: [SUN-SEG-Easy](https://paperswithcode.com/dataset/sun-seg-easy) & [SUN-SEG-Hard](https://paperswithcode.com/dataset/sun-seg-hard).
Here, we present a variety of qualitative and quantitative results of VPS benchmarks:
- Visual prediction of top-performance competitors:
<p align="center">
<img src="./assets/Qual-min.gif"/> <br />
<em>
Figure 3: Qualitative comparison of three video-based models (PNS+, PNSNet, and 2/3D) and two image-based models (ACSNet, and PraNet).
</em>
</p>
- Model-based performance:
<p align="center">
<img src="./assets/ModelPerformance.png"/> <br />
<em>
Figure 4: Quantitative comparison on two testing sub-datasets, i.e., SUN-SEG-Easy (Unseen) and SUN-SEG-Hard (Unseen). `R/T' represents we re-train the non-public model, whose code is provided by the original authors. The best scores are highlighted in bold.
</em>
</p>
- Attribute-based performance:
<p align="center">
<img src="./assets/AttributePerformance.png"/> <br />
<em>
Figure 5: Visual attributes-based performance on our SUN-SEG-Easy (Unseen) and SUN-SEG-Hard (Unseen) in terms of structure measure.
</em>
</p>
# 6. Tracking Trends
<p align="center">
<img src="./assets/the-reading-list.png"/> <br />
</p>
To better understand the development of this field and to quickly push researchers in their research process, we elaborately build a **Paper Reading List**. It includes **119** colonoscopy imaging-based AI scientific research in the past 12 years. It includes several fields, such as image polyp segmentation, video polyp segmentation, image polyp detection, video polyp detection, and image polyp classification. Besides, we will provide some interesting resources about human colonoscopy.
> **Note:** If we miss some treasure works, please let me know via e-mail or directly push a PR. We will work on it as soon as possible. Many thanks for your active feedback.
- The latest paper reading list and some interesting resources refer to [`Awesome-Video-Polyp-Segmentation.md`](https://github.com/GewelsJI/VPS/blob/main/docs/AWESOME_VPS.md)
# 7. Citations
If you have found our work useful, please use the following reference to cite this project:
@article{ji2022video,
title={Video polyp segmentation: A deep learning perspective},
author={Ji, Ge-Peng and Xiao, Guobao and Chou, Yu-Cheng and Fan, Deng-Ping and Zhao, Kai and Chen, Geng and Van Gool, Luc},
journal={Machine Intelligence Research},
volume={19},
number={6},
pages={531--549},
year={2022},
publisher={Springer}
}
@inproceedings{ji2021progressively,
title={Progressively normalized self-attention network for video polyp segmentation},
author={Ji, Ge-Peng and Chou, Yu-Cheng and Fan, Deng-Ping and Chen, Geng and Fu, Huazhu and Jha, Debesh and Shao, Ling},
booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
pages={142--152},
year={2021},
organization={Springer}
}
@inproceedings{fan2020pranet,
title={Pranet: Parallel reverse attention network for polyp segmentation},
author={Fan, Deng-Ping and Ji, Ge-Peng and Zhou, Tao and Chen, Geng and Fu, Huazhu and Shen, Jianbing and Shao, Ling},
booktitle={International conference on medical image computing and computer-assisted intervention},
pages={263--273},
year={2020},
organization={Springer}
}
# 8. FAQ
- Thanks to [Tuo Wang](victor_wt@qq.com) for providing a great solution to [upgrade the CUDA version when compiling the NS block](./docs/Upgrade%20environment%20for%20NS%20block.pdf).
# 9. License
The dataset and source code is free for research and education use only. Any commercial usage should get formal permission first.
- **Video Source:** SUN (Showa University and Nagoya University) Colonoscopy Video Database is the colonoscopy video database for the evaluation of automated colorectal-polyp detection. The database comprises still images of videos, which are collected at the Showa University Northern Yokohama Hospital. Mori Laboratory, Graduate School of Informatics, Nagoya University developed this database. Every frame in the database was annotated by the expert endoscopists at Showa University.
- **Intended Use:** This database is available for only non-commercial use in research or educational purposes.
As long as you use the database for these purposes, you can edit or process images and annotations in this database.
Without permission from Mori Lab., commercial use of this dataset is prohibited even after copying, editing,
processing, or any operations of this database. Please contact us for commercial use or if you are uncertain about
the decision.
- **Distribution:** It is prohibited to sell, transfer, lend, lease, resell, distribute, etc., as it is, or copy, edit, or process this database, in whole or in part.
# 10. Acknowledgments
- Our dataset is built upon the SUN (Showa University and Nagoya University) Colonoscopy Video Database, Thanks very much for their wonderful work!
- This codebase is based on our conference version [PNSNet](https://github.com/GewelsJI/PNS-Net), which is accepted by the MICCAI-2021 conference.
================================================
FILE: docs/AWESOME_VPS.md
================================================
# Awesome Video Polyp Segmentation
<img src="https://img.shields.io/badge/Contributions-Welcome-278ea5" alt="Contrib"/>
<img src="https://img.shields.io/badge/Number%20of%20Papers-192-FF6F00" alt="PaperNum"/>
# 1. Preview
This is a paper collection of **133** colonoscopy imaging-based AI scientific researches in recent **12** years.
In order to better understand the development of this field and to help researchers in their research process, we have divided the works into five tasks, including **133** papers on [image polyp segmentation](#21-image-polyp-segmentation), **5** papers on [video polyp segmentation](#22-video-polyp-segmentation), **17** papers on [image polyp detection](#23-image-polyp-detection), **11** papers on [video polyp detection](#24-video-polyp-detection), **6** paper on [image polyp classification](#25-image-polyp-classification), **1** paper on [video polyp classification](#26-video-polyp-classification), **2** paper on [colonoscopy depth estimation](#27-colonoscopy-depth-estimation), **1** paper on [colonoscopy deficient coverage detection](#28-colonoscopy-deficient-coverage-detection), and **1** paper on [colon polyp image synthesis](#29-colon-polyp-image-synthesis).
Besides, we present the collection of **14** polyp related datasets, including **8** [image segmentation datasets](#31-image-segmentation-datasets), **2** [video segmentation datasets](#32-video-segmentation-datasets), **1** [video detection dataset](#33-video-detection-datasets), **3** [video classification datasets](#34-video-classification-datasets), and **2** [colonoscopy depth dataset](#35-colonoscopy-depth-datasets).
In addition, we provide links to each paper and its repository whenever possible. * denotes the corresponding paper cannot be downloaded or the link is connected to the journal.
> Note that this page is under construction. If you have anything to recommend or any suggestions, please feel free to contact us via e-mail (gepengai.ji@gmail) or directly push a PR.
--- *Last updated: 12/07/2022* ---
## 1.1. Table of Contents
- [Awesome Video Polyp Segmentation](#awesome-video-polyp-segmentation)
- [1. Preview](#1-preview)
* [1.1. Table of Contents](#11-table-of-contents)
- [2. Polyp Related Methods](#2-polyp-related-methods)
* [2.1 Image Polyp Segmentation](#21-image-polyp-segmentation)
* [2.2 Video Polyp Segmentation](#22-video-polyp-segmentation)
* [2.3 Image Polyp Detection](#23-image-polyp-detection)
* [2.4 Video Polyp Detection](#24-video-polyp-detection)
* [2.5 Image Polyp Classification](#25-image-polyp-classification)
* [2.6 Video Polyp Classification](#26-video-polyp-classification)
* [2.7 Colonoscopy Depth Estimation](#27-colonoscopy-depth-estimation)
* [2.8 Colonoscopy Deficient Coverage Detection](#28-colonoscopy-deficient-coverage-detection)
* [2.9 Colon Polyp Image Synthesis](#29-colon-polyp-image-synthesis)
- [3. Polyp Related Datasets](#3-useful-resources)
* [3.1 Image Segmentation Datasets](#31-image-segmentation-datasets)
* [3.2 Video Segmentation Datasets](#32-video-segmentation-datasets)
* [3.3 Video Detection Datasets](#33-video-detection-datasets)
* [3.4 Video Classification Datasets](#34-video-classification-datasets)
* [3.5 Colonoscopy Depth Datasets](#35-colonoscopy-depth-datasets)
- [4. Useful Resources](#4-useful-resources)
* [4.1 Colonoscopy Related](#41-colonoscopy-related)
* [4.2 AI Conference Deadlines](#42-ai-conference-deadlines)
# 2. Polyp Related Methods
## 2.1 Image Polyp Segmentation
**Yr.** | **Pub.** | **Title** | **Links**
:-: | :-: | :- | :-:
2023 | **MICCAI** | WeakPolyp: You Only Look Bounding Box for Polyp Segmentation | [Paper](https://arxiv.org/pdf/2307.10912.pdf)/[Code](https://github.com/weijun88/WeakPolyp)
2023 | **MICCAI** | Revisiting Feature Propagation and Aggregation in Polyp Segmentation | Paper
2023 | **MICCAI** | Probabilistic Modeling Ensemble Vision Transformer Improves Complex Polyp Segmentation | Paper
2023 | **MICCAI** | S2ME: Spatial-Spectral Mutual Teaching and Ensemble Learning for Scribble-supervised Polyp Segmentation | [Paper](https://arxiv.org/abs/2306.00451)/[Code](https://github.com/lofrienger/S2ME)
2023 | **arXiv** | Can SAM Segment Polyps? | [Paper](https://arxiv.org/abs/2304.07583)
2023 | **arXiv** | Polyp-SAM++: Can A Text Guided SAM Perform Better for Polyp Segmentation? | [Paper](https://www.researchgate.net/publication/373091859_Polyp-SAM_Can_A_Text_Guided_SAM_Perform_Better_for_Polyp_Segmentation)
2023 | **arXiv** | Polyp-SAM: Transfer SAM for Polyp Segmentation | [Paper](https://arxiv.org/abs/2305.00293)
2023 | **PR** | Cross-level Feature Aggregation Network for Polyp Segmentation | [Paper](https://www.sciencedirect.com/science/article/pii/S0031320323002558)/[Code](https://github.com/taozh2017/CFANet)
2023 | **BMVC** | `ADSNet` ADSNet: Adaptation of Distinct Semantic for Uncertain Areas in Polyp Segmentation | [Paper](https://papers.bmvc2023.org/0806.pdf)
2022 | **ISICT** | Incremental Boundary Refinement using Self Axial Reverse Attention and Uncertainty-aware Gate for Colon Polyp Segmentation | [Paper*](https://dl.acm.org/doi/abs/10.1145/3568562.3568663)/Code
2022 | **TVCJ** | DCANet: deep context attention network for automatic polyp segmentation | [Paper](https://link.springer.com/content/pdf/10.1007/s00371-022-02677-x.pdf?pdf=button)/Code
2022 | **arXiv** | Towards Automated Polyp Segmentation Using Weakly- and Semi-Supervised Learning and Deformable Transformers | [Paper](https://arxiv.org/pdf/2211.11847.pdf)/Code
2022 | **arXiv** | Spatially Exclusive Pasting: A General Data Augmentation for the Polyp Segmentation | [Paper](https://arxiv.org/pdf/2211.08284.pdf)/Code
2022 | **CBM** | MSRAformer: Multiscale spatial reverse attention network for polyp segmentation | [Paper*](https://www.sciencedirect.com/science/article/abs/pii/S0010482522009829)/Code
2022 | **CBM** | DBMF: Dual Branch Multiscale Feature Fusion Network for polyp segmentation | [Paper*](https://www.sciencedirect.com/science/article/abs/pii/S0010482522010125)/Code
2022 | **ICAISM** | Automatic Polyp Segmentation in Colonoscopy Images Using Single Network Model: SegNet | [Paper](https://link.springer.com/content/pdf/10.1007/978-981-16-2183-3_69.pdf?pdf=inline%20link)/Code
2022 | **ICMAR** | Polyp segmentation algorithm combining multi-scale attention and multi-layer loss | [Paper*](https://www.spiedigitallibrary.org/conference-proceedings-of-spie/12331/123314W/Polyp-segmentation-algorithm-combining-multi-scale-attention-and-multi-layer/10.1117/12.2652907.short?SSO=1)/Code
2022 | **MICCAI** | Using Guided Self-Attention with Local Information for Polyp Segmentation | [Paper](https://link.springer.com/chapter/10.1007/978-3-031-16440-8_60)/Code
2022 | **MICCAI** | Task-Relevant Feature Replenishment for Cross-Centre Polyp Segmentation| [Paper](https://link.springer.com/chapter/10.1007/978-3-031-16440-8_57)/[Code](https://github.com/CathyS1996/TRFRNet)
2022 | **MICCAI** | TGANet: Text-guided attention for improved polyp segmentation | [Paper](https://arxiv.org/pdf/2205.04280.pdf)/[Code](https://github.com/nikhilroxtomar/TGANet)
2022 | **MICCAI** | Lesion-Aware Dynamic Kernel for Polyp Segmentation | [Paper](https://link.springer.com/chapter/10.1007/978-3-031-16437-8_10)/Code
2022 | **MICCAI** | Semi-Supervised Spatial Temporal Attention Network for Video Polyp Segmentation | [Paper](https://link.springer.com/chapter/10.1007/978-3-031-16437-8_7)/Code
2022 | **CMIG** | Boosting medical image segmentation via conditional-synergistic convolution and lesion decoupling | [Paper](https://www.sciencedirect.com/science/article/abs/pii/S0895611122000817)/[Code](https://github.com/QianChen98/CCLD-Net)
2022 | **Gastroenterology Insights** | UPolySeg: A U-Net-Based Polyp Segmentation Network Using Colonoscopy Images | [Paper](https://www.mdpi.com/2036-7422/13/3/27/pdf?version=1660117945)/Code
2022 | **Electronics** | A Segmentation Algorithm of Colonoscopy Images Based on Multi-Scale Feature Fusion | [Paper](https://www.mdpi.com/2079-9292/11/16/2501/pdf?version=1660199760)/Code
2022 | **TCSVT** | Polyp-Mixer: An Efficient Context-Aware MLP-based Paradigm for Polyp Segmentation | [Paper](https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9852486)/[Code](https://github.com/shijinghuihub/Polyp-Mixer)
2022 | **TETCI** | Adaptive Context Exploration Network for Polyp Segmentation in Colonoscopy Images | [Paper](https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9852746)/Code
2022 | **IJCAI** | ICGNet: Integration Context-Based Reverse-Contour Guidance Network for Polyp Segmentation | [Paper](https://www.ijcai.org/proceedings/2022/0123.pdf)/Code
2022 | **IJCAI** | TCCNet: Temporally Consistent Context-Free Network for Semi-supervised Video Polyp Segmentation | [Paper](https://www.ijcai.org/proceedings/2022/0155.pdf)/[Code](https://github.com/wener-yung/TCCNet)
2022 | **AIM** | An end-to-end tracking method for polyp detectors in colonoscopy videos | [Paper](https://reader.elsevier.com/reader/sd/pii/S0933365722001270?token=85406D788AF1B59597BD0BCA3456A70DD2ECE3040EBCB1C3D669584B712F58FA656224415F80070007005E3D36B81F8D&originRegion=us-east-1&originCreation=20220801152429)/Code
2022 | **MIUA** | Polyp2Seg: Improved Polyp Segmentation with Vision Transformer | [Paper](https://link.springer.com/content/pdf/10.1007/978-3-031-12053-4_39.pdf)/Code
2022 | **CVIP** | Localization of Polyps in WCE Images Using Deep Learning Segmentation Methods: A Comparative Study | [Paper](https://link.springer.com/content/pdf/10.1007/978-3-031-11346-8_46.pdf)/Code
2022 | **AIMD** | SARM-Net: A Spatial Attention-Based Residual M-Net for Polyp Segmentation | [Paper](https://link.springer.com/chapter/10.1007/978-981-19-0151-5_33)/Code
2022 | **BSPC** | FAPN: Feature Augmented Pyramid Network for polyp segmentation | [Paper](https://reader.elsevier.com/reader/sd/pii/S1746809422004074?token=FE011B0123F802F27442369ED87DD656B402B1920147F48DC6957C72D5D1859DFD0B1ADB80194C54FBB05A9220781C20&originRegion=us-east-1&originCreation=20220708075935)/Code
2022 | **BSPC** | Automated polyp segmentation in colonoscopy images via deep network with lesion-aware feature selection and refinement | [Paper](https://www.sciencedirect.com/sdfe/reader/pii/S1746809422003688/pdf)/Code
2022 | **ICFCS** | U-Shaped Xception-Residual Network for Polyps Region Segmentation | [Paper](https://link.springer.com/chapter/10.1007/978-981-19-0105-8_25)/Code
2022 | **CBM** | Colorectal polyp region extraction using saliency detection network with neutrosophic enhancement | [Paper](https://reader.elsevier.com/reader/sd/pii/S0010482522005340?token=0E8C163715B86CD6B6EC6F0D60685AA70877424A885FF787BB2C776923BB44A84AF65684FBCAC9503E28CA794B58174B&originRegion=us-east-1&originCreation=20220708080248)/Code
2022 | **IJCARS** | Examining the effect of synthetic data augmentation in polyp detection and segmentation | [Paper](https://link.springer.com/content/pdf/10.1007/s11548-022-02651-x.pdf)/Code
2022 | **IEEE Access** | Polyp Segmentation of Colonoscopy Images by Exploring the Uncertain Areas | [Paper](https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9775966)/Code
2022 | **JBHI** | Boundary Constraint Network with Cross Layer Feature Integration for Polyp Segmentation | [Paper](https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9772424)/Code
2022 | **CMIG** | Polyp Segmentation Network with Hybrid Channel-Spatial Attention and Pyramid Global Context Guided Feature Fusion | [Paper](https://reader.elsevier.com/reader/sd/pii/S0895611122000453?token=1CC9A6070522894EA16E7984593F634B2A32CCB11CB9CBFD4CAA37916A13DE4D3F6AB47DFDDB5F6ED12F23B0CAD0FE20&originRegion=us-east-1&originCreation=20220515094414)/Code
2022 | **arXiv** | Automatic Polyp Segmentation with Multiple Kernel Dilated Convolution Network | [Paper](https://arxiv.org/pdf/2206.06264.pdf)/Code
2022 | **arXiv** | PlutoNet: An Efficient Polyp Segmentation Network | [Paper](https://arxiv.org/pdf/2204.03652.pdf)/Code
2022 | **arXiv** | Automated Polyp Segmentation in Colonoscopy using MSRFNet | [Paper](https://www.researchgate.net/profile/Debesh-Jha/publication/359698512_Automated_Polyp_Segmentation_in_Colonoscopy_using_MSRFNet/links/624907bf8068956f3c6533c1/Automated-Polyp-Segmentation-in-Colonoscopy-using-MSRFNet.pdf)/Code
2022 | **arXiv** | BlazeNeo: Blazing fast polyp segmentation and neoplasm detection | [Paper](https://arxiv.org/pdf/2203.00129.pdf)/Code
2022 | **arXiv** | BDG-Net: Boundary Distribution Guided Network for Accurate Polyp Segmentation | [Paper](https://arxiv.org/pdf/2201.00767.pdf)/Code
2022 | **arXiv** | Cross-level Contrastive Learning and Consistency Constraint for Semi-supervised Medical Image Segmentation | [Paper](https://arxiv.org/pdf/2202.04074.pdf)/Code
2022 | **arXiv** | ColonFormer: An Efficient Transformer based Method for Colon Polyp Segmentation | [Paper](https://arxiv.org/pdf/2205.08473.pdf)/Code
2022 | **KBS** | MIA-Net: Multi-information aggregation network combining transformers and convolutional feature learning for polyp segmentation | [Paper](https://www.sciencedirect.com/science/article/pii/S0950705122003926)/Code
2022 | **Diagnostics** | Performance of Convolutional Neural Networks for Polyp Localization on Public Colonoscopy Image Datasets | [Paper](https://www.mdpi.com/2075-4418/12/4/898/htm)/Code
2022 | **IEEE TCyber** | PolypSeg+: A Lightweight Context-Aware Network for Real-Time Polyp Segmentation | [Paper](https://ieeexplore.ieee.org/abstract/document/9756512)/Code
2022 | **JCDE** | SwinE-Net: hybrid deep learning approach to novel polyp segmentation using convolutional neural network and Swin Transformer | [Paper](https://academic.oup.com/jcde/article/9/2/616/6564811?login=true)/Code
2022 | **IEEE JBHI** | Artificial Intelligence for Colonoscopy: Past, Present, and Future | [Paper](https://ieeexplore.ieee.org/document/9739863)/Code
2021 | **ICONIP** | Multi-scale Fusion Attention Network for Polyp Segmentation | [Paper](https://link.springer.com/chapter/10.1007/978-3-030-92310-5_19)/Code
2021 | **AAAI** | Precise yet Efficient Semantic Calibration and Refinement in ConvNets for Real-time Polyp Segmentation from Colonoscopy Videos | [Paper](https://www.aaai.org/AAAI21Papers/AAAI-5002.WuHS.pdf)/Code
2021 | **ICCV** | Collaborative and Adversarial Learning of Focused and Dispersive Representations for Semi-supervised Polyp Segmentation | [Paper](https://openaccess.thecvf.com/content/ICCV2021/papers/Wu_Collaborative_and_Adversarial_Learning_of_Focused_and_Dispersive_Representations_for_ICCV_2021_paper.pdf)/Code
2021 | **ACM MM** | UACANet: Uncertainty Augmented Context Attention for Polyp Segmentation | [Paper](https://dl.acm.org/doi/pdf/10.1145/3474085.3475375)/[Code](https://github.com/plemeri/UACANet)
2021 | **Healthcare** | TMD-Unet: Triple-Unet with Multi-Scale Input Features and Dense Skip Connection for Medical Image Segmentation | [Paper](https://www.researchgate.net/publication/348283572_TMD-Unet_Triple-Unet_with_Multi-Scale_Input_Features_and_Dense_Skip_Connection_for_Medical_Image_Segmentation)/Code
2021 | **ICPR** | DDANet: Dual decoder attention network for automatic polyp segmentation | [Paper](https://arxiv.org/pdf/2012.15245.pdf)/[Code](https://github.com/nikhilroxtomar/DDANet)
2021 | **ICDSIT** | Sa-HarDNeSt: A Self-Attention Network for Polyp Segmentation | [Paper](https://dl.acm.org/doi/pdf/10.1145/3478905.3478942)/Code
2021 | **IJCAI** | Medical image segmentation using squeeze-and-expansion transformers | [Paper](https://arxiv.org/pdf/2105.09511.pdf)/[Code](https://github.com/askerlee/segtran)
2021 | **IEEE ISBI** | DivergentNets: Medical Image Segmentation by Network Ensemble | [Paper](https://arxiv.org/pdf/2107.00283.pdf)/[Code](https://github.com/vlbthambawita/divergent-nets)
2021 | **IEEE JBHI** | A comprehensive study on colorectal polyp segmentation with resunet++, conditional random field and test-time augmentation | [Paper](https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9314114)/[Code](https://github.com/DebeshJha/ResUNetPlusPlus-with-CRF-and-TTA)
2021 | **IEEE JBHI** | Mutual-prototype adaptation for cross-domain polyp segmentation | [Paper](https://sci-hub.se/downloads/2021-05-24/1a/yang2021.pdf?rand=61ab86a0abb28?download=true)/[Code](https://github.com/CityU-AIM-Group/MPA-DA)
2021 | **MIA** | Dynamic-weighting hierarchical segmentation network for medical images | [Paper](https://pdf.sciencedirectassets.com/272154/1-s2.0-S1361841521X00059/1-s2.0-S1361841521002413/main.pdf?X-Amz-Security-Token=IQoJb3JpZ2luX2VjEIf%2F%2F%2F%2F%2F%2F%2F%2F%2F%2FwEaCXVzLWVhc3QtMSJHMEUCIQDZ0ITNVi%2BvzuLOUKPQQajvIwWt9hsidNnvAiweckL5rgIgQv2G7DZIc6bMCiIaipFlvaKM8zJWKR%2BJ75Q6tnwBWVQq%2BgMIYBAEGgwwNTkwMDM1NDY4NjUiDMa65PEpU2yP8TutHirXA2%2FvwxG2Px%2B5huEgdCPa%2BWft55sBiAK71F3Ebz%2Fj7AJ2EHzmyFNE4%2BtBAHS%2Ft2dAE0l9X1a8DBEj2hj%2BnQfo5lfkj1bS6gxEny5IHEKozs9X%2BAt1l1rv7PIPXN6Eb6%2B5%2FQe24O%2Bu6iJyeiTbUS2Pk3kZCiyIbVNRygfDn6j8l5Ye1JqLSl8zljMiZJBZKWfE9pekhQbKPi5pyqflJmiBhZFxuI2YGjOhQ0LhY0fQKIqrAu0AWvFXBdv%2BX%2FxLHBStP911TVbrCl6zCf4m1Y%2FYmrRkEiR343YY0VuJ%2Bswg8p23Lf7nQ3tcSvKaJ5WmpINOPl3O%2BA7AdiqOkvF2%2BklOIfqpNSn1kA591KHwjYh4tSIQdlVCBYQIOFb5DGD3hfnwUwery0DdMRC0H4vgChsyEPGUiI2JVgr6WytCvKl3%2Bnm%2FlaxUopskmpT7S5QGGIYkKy56VoBL1yBP4SqlU1bWe%2FbujaO6mq6i5xjPP8W2l4OlCZIwKX79dq17AIN%2B1cZZYp84zVSjYpyn8qSXkxCnf3x8UNKws2TzALIMQl9YQKloABt%2BvLQYoN7P6MINo%2FcB%2FaC6HP3yXMuho7oorZtPgU5P3IHpaHSxod3iKD5uvh87P2h2noQt%2FTCQ9K2NBjqlAQj%2FqBYVx%2BDjQ%2BXuOFSdHTIw2tRNrxOxQsEtmxdxy7ej5ttMEQKnq3rhA1K6ETIoS5IH5RNNgGzijDwYiNGXy4tJd7k3tmoFnoqicAqyycG%2FPu2gzvYGKhVNh9bxllTdijLBfJRTMKSMFwyH1W3JqFZUDbT9qO7bwqciDPMerP6zF0KdGfjBrY%2FPWYJf%2FWOdhSBBTE62hfm%2FW573OuLkI1KwU1dWrw%3D%3D&X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Date=20211204T153800Z&X-Amz-SignedHeaders=host&X-Amz-Expires=300&X-Amz-Credential=ASIAQ3PHCVTY4BAUX3ZO%2F20211204%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Signature=80ffde177a66c02f8f70db6588324bca239c48bf4366af906e2cf876c11dbcea&hash=854aafe64938f04f8ee0bf386f4c92d77219c66c67bad994e7d8806082bdafd1&host=68042c943591013ac2b2430a89b270f6af2c76d8dfd086a07176afe7c76c2c61&pii=S1361841521002413&tid=spdf-875a5c1f-84ac-43d9-aff2-8406cae898df&sid=ceb35d509987134d700962537e5a5b60cfe0gxrqa&type=client)/[Code](https://github.com/CityU-AIM-Group/DW-HieraSeg)
2021 | **MICCAI** | Automatic Polyp Segmentation via Multi-scale Subtraction Network | [Paper](https://arxiv.org/pdf/2108.05082.pdf)/[Code](https://github.com/Xiaoqi-Zhao-DLUT/MSNet)
2021 | **MICCAI** | CCBANet: Cascading Context and Balancing Attention for Polyp Segmentation | [Paper*](https://link.springer.com/content/pdf/10.1007%2F978-3-030-87193-2.pdf)/[Code](https://github.com/ntcongvn/CCBANet)
2021 | **MICCAI** | Constrained Contrastive Distribution Learning for Unsupervised Anomaly Detection and Localisation in Medical Images | [Paper](https://arxiv.org/pdf/2103.03423.pdf)/[Code](https://arxiv.org/pdf/2103.03423.pdf)
2021 | **MICCAI** | Double Encoder-Decoder Networks for Gastrointestinal Polyp Segmentation | [Paper](https://arxiv.org/pdf/2110.01939.pdf)/Code
2021 | **MICCAI** | HRENet: A Hard Region Enhancement Network for Polyp Segmentation | [Paper*](https://link.springer.com/content/pdf/10.1007%2F978-3-030-87193-2.pdf)/[Code](https://github.com/CathySH/HRENet)
2021 | **MICCAI** | Few-Shot Domain Adaptation with Polymorphic Transformers | [Paper](https://arxiv.org/pdf/2107.04805.pdf)/[Code](https://github.com/askerlee/segtran)
2021 | **MICCAI** | Learnable Oriented-Derivative Network for Polyp Segmentation | [Paper*](https://link.springer.com/content/pdf/10.1007%2F978-3-030-87193-2.pdf)/[Code](https://github.com/midsdsy/LOD-Net)
2021 | **MICCAI** | Shallow attention network for polyp segmentation | [Paper](https://arxiv.org/pdf/2108.00882.pdf)/[Code](https://github.com/weijun88/SANet)
2021 | **MICCAI** | Transfuse: Fusing transformers and cnns for medical image segmentation | [Paper](https://arxiv.org/pdf/2102.08005.pdf)/Code
2021 | **MIDL** | Deep ensembles based on stochastic activation selection for polyp segmentation | [Paper](https://arxiv.org/pdf/2104.00850.pdf)/[Code](https://github.com/LorisNanni/Deep-ensembles-based-on-Stochastic-Activation-Selection-for-Polyp-Segmentation)
2021 | **NCBI** | MBFFNet: Multi-Branch Feature Fusion Network for Colonoscopy | [Paper](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8317500/pdf/fbioe-09-696251.pdf)/Code
2021 | **RIVF** | AG-CUResNeSt: A novel method for colon polyp segmentation | [Paper](https://arxiv.org/pdf/2105.00402.pdf)/Code
2021 | **Sensors** | A-DenseUNet: Adaptive Densely Connected UNet for Polyp Segmentation in Colonoscopy Images with Atrous Convolution | [Paper](https://www.mdpi.com/1424-8220/21/4/1441/pdf)/Code
2021 | **IEEE TIM** | Colon Polyp Detection and Segmentation Based on Improved MRCNN | [Paper](https://www.researchgate.net/publication/346985142_Colon_Polyp_Detection_and_Segmentation_based_on_improved_MRCNN)/Code
2021 | **IEEE TIM** | Polyp-Net A Multimodel Fusion Network for Polyp Segmentation | [Paper](https://drive.google.com/file/d/1isi_Blz9ZAK4iPH5wKcEVw4-FSYuqNxm/view)/Code
2021 | **IEEE TMI** | Graph-based Region and Boundary Aggregation for Biomedical Image Segmentation | [Paper](https://livrepository.liverpool.ac.uk/3140502/1/TMI_region_boundary2021.pdf)/[Code](https://github.com/smallmax00/Graph_Region_Boudnary)
2021 | **IEEE Access** | A Simple Generic Method for Effective Boundary Extraction in Medical Image Segmentation | [Paper*](https://ieeexplore.ieee.org/iel7/6287639/9312710/09495769.pdf)/Code
2021 | **IEEE Access** | CRF-EfficientUNet: An Improved UNet Framework for Polyp Segmentation in Colonoscopy Images With Combined Asymmetric Loss Function and CRF-RNN Layer | [Paper](https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9622208)/[Code](https://github.com/lethithuhong1302/CRF-EfficientUNet)
2021 | **IEEE Access** | Real-time polyp detection, localization and segmentation in colonoscopy using deep learning | [Paper](https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9369308)/[Code](https://github.com/DebeshJha/ColonSegNet)
2021 | **IEEE Access** | Training on Polar Image Transformations Improves Biomedical Image Segmentation | [Paper](https://ieeexplore.ieee.org/iel7/6287639/9312710/09551998.pdf)/Code
2021 | **BioMed** | Automated Classification and Segmentation in Colorectal Images Based on Self-Paced Transfer Network | [Paper](https://www.hindawi.com/journals/bmri/2021/6683931/)/Code
2021 | **CBM** | Focus U-Net: A novel dual attention-gated CNN for polyp segmentation during colonoscopy | [Paper](https://sci-hub.se/10.1016/j.compbiomed.2021.104815)/Code
2021 | **CBMS** | Nanonet: Real-time polyp segmentation in video capsule endoscopy and colonoscopy | [Paper](https://arxiv.org/pdf/2104.11138.pdf)/[Code](https://github.com/DebeshJha/NanoNet)
2021 | **CRV** | Enhanced u-net: A feature enhancement network for polyp segmentation | [Paper](https://arxiv.org/pdf/2105.00999.pdf)/[Code](https://github.com/rucv/Enhanced-U-Net)
2021 | **IEEE DDCLS** | MSB-Net: Multi-Scale Boundary Net for Polyp Segmentation | [Paper*](https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9455514)/Code
2021 | **arXiv** | BI-GCN: Boundary-Aware Input-Dependent Graph Convolution Network for Biomedical Image Segmentation | [Paper](https://arxiv.org/pdf/2110.14775.pdf)/Code
2021 | **arXiv** | CaraNet: Context Axial Reverse Attention Network for Segmentation of Small Medical Objects | [Paper](https://arxiv.org/pdf/2108.07368.pdf)/Code
2021 | **arXiv** | DS-TransUNet: Dual Swin Transformer U-Net for Medical Image Segmentation | [Paper](https://arxiv.org/pdf/2106.06716.pdf)/Code
2021 | **arXiv** | Duplex contextual relation network for polyp segmentation | [Paper](https://arxiv.org/pdf/2103.06725)/[Code](https://github.com/PRIS-CV/DCRNet)
2021 | **arXiv** | Few-shot segmentation of medical images based on meta-learning with implicit gradients | [Paper](https://arxiv.org/pdf/2106.03223.pdf)/Code
2021 | **arXiv** | GMSRF-Net: An improved generalizability with global multi-scale residual fusion network for polyp segmentation | [Paper](https://arxiv.org/pdf/2111.10614.pdf)/Code
2021 | **arXiv** | Hardnet-mseg: A simple encoder-decoder polyp segmentation neural network that achieves over 0.9 mean dice and 86 fps | [Paper](https://arxiv.org/pdf/2101.07172.pdf)/[Code](https://github.com/james128333/HarDNet-MSEG)
2021 | **arXiv** | NeoUNet: Towards accurate colon polyp segmentation and neoplasm detection | [Paper](https://arxiv.org/pdf/2107.05023.pdf)/Code
2021 | **arXiv** | Polyp segmentation in colonoscopy images using u-net-mobilenetv2 | [Paper](https://arxiv.org/pdf/2103.15715.pdf)/Code
2021 | **arXiv** | Polyp-PVT: Polyp Segmentation with Pyramid Vision Transformers | [Paper](https://arxiv.org/pdf/2108.06932.pdf)/[Code](https://github.com/DengPingFan/Polyp-PVT)
2021 | **arXiv** | Self-supervised Multi-class Pre-training for Unsupervised Anomaly Detection and Segmentation in Medical Images | [Paper](https://arxiv.org/pdf/2109.01303.pdf)/Code
2020 | **IEEE Access** | Contour-Aware Polyp Segmentation in Colonoscopy Images Using Detailed Upsamling Encoder-Decoder Networks | [Paper](https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9096362)/Code
2020 | **arXiv** | Automatic polyp segmentation using convolution neural networks | [Paper](https://arxiv.org/pdf/2004.10792)/Code
2020 | **arXiv** | Boundary-aware Context Neural Network for Medical Image Segmentation | [Paper](https://arxiv.org/pdf/2005.00966v1.pdf)/Code
2020 | **CBMS** | DoubleU-Net A Deep Convolutional Neural Network for Medical Image Segmentation | [Paper](https://arxiv.org/pdf/2006.04868)/[Code](https://github.com/DebeshJha/2020-CBMS-DoubleU-Net)
2020 | **HYDCON** | Polyps Segmentation using Fuzzy Thresholding in HSV Color Space | [Paper](https://sci-hub.se/downloads/2020-12-21/80/mandal2020.pdf?rand=61ab7ececa47d?download=true)/Code
2020 | **ICARM** | Real-time Colonoscopy Image Segmentation Based on Ensemble Knowledge Distillation | [Paper](https://sci-hub.se/downloads/2020-11-09/c2/huang2020.pdf?rand=61ab7ea22a48f?download=true)/Code
2020 | **IEEE ISBI** | SSN A stair-shape network for real-time polyp segmentation in colonoscopy images | [Paper](https://www.researchgate.net/profile/Ruiwei-Feng/publication/339782695_SSN_A_Stair-Shape_Network_for_Real-Time_Polyp_Segmentation_in_Colonoscopy_Images/links/5e6af8dd458515e555765049/SSN-A-Stair-Shape-Network-for-Real-Time-Polyp-Segmentation-in-Colonoscopy-Images.pdf)/Code
2020 | **IEEE JBHI** | Multi-scale Context-guided Deep Network for Automated Lesion Segmentation with Endoscopy Images of Gastrointestinal Tract | [Paper](https://sci-hub.se/downloads/2020-06-18//db/wang2020.pdf?rand=61ab7f237a9b5?download=true)/Code
2020 | **MIA** | Uncertainty and interpretability in convolutional neural networks for semantic segmentation of colorectal polyps | [Paper](https://pdf.sciencedirectassets.com/272154/1-s2.0-S1361841519X00092/1-s2.0-S1361841519301574/main.pdf?X-Amz-Security-Token=IQoJb3JpZ2luX2VjEKv%2F%2F%2F%2F%2F%2F%2F%2F%2F%2FwEaCXVzLWVhc3QtMSJGMEQCIFCBk%2FghOtJfQ62sxbB63ru7AOYC0IBJiHp%2F3HJrjaU%2BAiANHFD06HdAPYve6VbMJLdwkGojYzHjBDAL846ZP7ss%2FiqDBAiE%2F%2F%2F%2F%2F%2F%2F%2F%2F%2F8BEAQaDDA1OTAwMzU0Njg2NSIMdDwgrrObNsP3CG8sKtcDt1QC6gBH9G9et33b26BRppTYQZooVUiDID3EvU6AAsmbNCFp5XTmqgUR5rK90BvBAFaKmWgIVzhR4Hzs2%2B1knkFJMtwbhLFSPjSZ6rx%2F4%2FNI2%2FyPeMY5xN3Qd1WdRfM4I2HF%2Bmy3aO2xX1j4IchIAPhE5FVuXDE1BqWefNjTLHxO9Or7i8FcFktna%2BIB1nwJVuXieWZNtyQhRwa%2BxRDo1TvyBZ3FXQ4UJ0wVQT6ndB6eBuKMXRJrDlBzg2MlKxIII65ufBi5GhVJLqY7lZCfFswH2DhBXPYvmnqi64jrhPq6i4iRMUu%2FU78b7ibopJgGeWbqg337XQppOTHxiGxll2dhvJ7PMb%2FtN7WTWIkEYGvPr%2FPCf9ccgo%2FJDJZi7n7u9h1%2BEqNU5YPgHmW6N%2FIK19qSQrUbKq35DMgajbzHtqOSUYww2g%2FeLn8Ymyty19TwfWC%2BWAJfT%2BgbJGt3c%2BhEudC2o%2B37%2BZ9dlAjFdGLNrV%2BEb2sTR6753%2FLynwAL2ateKKWQkTrurKfmPIdVnj0J7mLMx9QXZI0nrVzjc8RJt%2FIGuDLHJbpgNjZcRPrV5G0d%2BfeO2n%2F3uy2Bxmh%2BQoKt%2F3m9xLE5SpVjjF5U%2FwjhAWQbZZPDdc8rMJ7otY0GOqYB%2Fg32ebRx9MGzVLDd%2B6k8bZYrtUqolCXWBBNThlnmcVMKVY%2FXxcRh2K1y5BGXKlDivpzB9dYmRfT3B81mJ3DG4OlS5n1voC8yiHIn0qleoeyvgph8qbuFlPJlxq51MXcHA1VZtvJYe5wTrznNCmZqP4Bby56JSVVdlzMrS%2FPhdDZBRjzqAzGO8nNvIeCgznTGUd1hpFqC5rjRCshoaxJG5V%2B3lR5qXg%3D%3D&X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Date=20211206T032453Z&X-Amz-SignedHeaders=host&X-Amz-Expires=300&X-Amz-Credential=ASIAQ3PHCVTYRZJU4SVY%2F20211206%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Signature=c5b143cc04624e9d899f887cf83cf56cbf27185fa0286d447629ed2a306a7612&hash=7bbe140b0a4e1e4a8c73f2b402ff854818267ec8f356b261b5b3cb4e78e96de0&host=68042c943591013ac2b2430a89b270f6af2c76d8dfd086a07176afe7c76c2c61&pii=S1361841519301574&tid=spdf-a7da0fef-5b1e-49c4-a75e-040ca7ca6680&sid=ec558a9b400130484b2a9f1-fbc4af95d4adgxrqa&type=client)/Code
2020 | **MICCAI** | Adaptive context selection for polyp segmentation | [Paper](https://www.sysuhcp.com/userfiles/files/2021/02/28/8323749c38f384d5.pdf)/[Code](https://github.com/ReaFly/ACSNet)
2020 | **MICCAI** | Mi2gan: Generative adversarial network for medical image domain adaptation using mutual information constraint | [Paper](https://arxiv.org/pdf/2007.11180.pdf)/Code
2020 | **MICCAI** | Pranet: Parallel reverse attention network for polyp segmentation | [Paper](https://arxiv.org/pdf/2006.11392.pdf)/[Code](https://github.com/DengPingFan/PraNet)
2020 | **MICCAI** | PolypSeg: An Efficient Context-Aware Network for Polyp Segmentation from Colonoscopy Videos | [Paper](https://link.springer.com/chapter/10.1007%2F978-3-030-59725-2_28)/Code
2020 | **MIUA** | Polyp Segmentation with Fully Convolutional Deep Dilation Neural Network | [Paper*](https://link.springer.com/content/pdf/10.1007%2F978-3-030-39343-4.pdf)/Code
2020 | **RIVF** | Polyp Segmentation in Colonoscopy Images Using Ensembles of U-Nets with EfficientNet and Asymmetric Similarity Loss Function | [Paper](http://eprints.uet.vnu.edu.vn/eprints/id/document/3713)/Code
2020 | **Sensors** | ABC-Net Area-Boundary Constraint Network with Dynamical Feature Selection for Colorectal Polyp Segmentation | [Paper](https://sci-hub.se/downloads/2020-08-26/00/fang2020.pdf?rand=61ab823b2d3f0?download=true)/Code
2020 | **IEEE TMI** | Learn to Threshold: Thresholdnet with confidence-guided manifold mixup for polyp segmentation | [Paper](https://www.researchgate.net/publication/347925189_Learn_to_Threshold_ThresholdNet_With_Confidence-Guided_Manifold_Mixup_for_Polyp_Segmentation)/[Code](https://github.com/Guo-Xiaoqing/ThresholdNet)
2019 | **IEEE Access** | Ensemble of Instance Segmentation Models for Polyp Segmentation in Colonoscopy Images | [Paper](https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=8648333)/Code
2019 | **CBMS** | Training Data Enhancements for Robust Polyp Segmentation in Colonoscopy Images | [Paper](https://webserver2.tecgraf.puc-rio.br/~abraposo/pubs/CBMS2019/08787526.pdf)/Code
2019 | **EMBC** | Psi-net: Shape and boundary aware joint multi-task deep network for medical image segmentation | [Paper](https://arxiv.org/pdf/1902.04099.pdf)/[Code](https://github.com/Bala93/Multi-task-deep-network)
2019 | **EMBC** | Polyp Segmentation using Generative Adversarial Network | [Paper*](https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8857958)/Code
2019 | **ICMLA** | Colorectal polyp segmentation by u-net with dilation convolution | [Paper](https://arxiv.org/pdf/1912.11947.pdf)/Code
2019 | **ISM** | ResUNet++: An Advanced Architecture for Medical Image Segmentation | [Paper](https://arxiv.org/pdf/1911.07067.pdf)/[Code](https://github.com/DebeshJha/ResUNetPlusPlus-with-CRF-and-TTA)
2019 | **ISMICT** | Polyp detection and segmentation using mask r-cnn: Does a deeper feature extractor cnn always perform better? | [Paper](https://arxiv.org/pdf/1907.09180.pdf)/Code
2019 | **MICCAI** | Selective Feature Aggregation Network with Area-Boundary Constraints for Polyp Segmentation | [Paper*](https://link.springer.com/content/pdf/10.1007%2F978-3-030-32239-7.pdf)/Code
2019 | **Nature** | U-Net – Deep Learning for Cell Counting, Detection, and Morphometry | [Paper](https://lmb.informatik.uni-freiburg.de/Publications/2019/FMBCAMBBR19/paper-U-Net.pdf)/[Code](https://lmb.informatik.uni-freiburg.de/resources/opensource/unet/)
2018 | **DLMIA** | Unet++: A nested u-net architecture for medical image segmentation | [Paper](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7329239/pdf/nihms-1600717.pdf)/[Code](https://github.com/MrGiovanni/UNetPlusPlus)
2018 | **EMBC** | Polyp segmentation in colonoscopy images using fully convolutional network | [Paper](https://arxiv.org/pdf/1802.00368.pdf)/Code
2018 | **ICMM** | Real-Time Polyps Segmentation for Colonoscopy Video Frames Using Compressed Fully Convolutional Network | [Paper](https://www.researchgate.net/profile/Can-Udomcharoenchaikit/publication/322424455_Real-Time_Polyps_Segmentation_for_Colonoscopy_Video_Frames_Using_Compressed_Fully_Convolutional_Network/links/5ecc209a299bf1c09adf5049/Real-Time-Polyps-Segmentation-for-Colonoscopy-Video-Frames-Using-Compressed-Fully-Convolutional-Network.pdf)/Code
2018 | **JMRR** | Towards a computed-aided diagnosis system in colonoscopy | [Paper](https://arxiv.org/pdf/2101.06040.pdf)/Code
2018 | **Medical Robotics Res** | Automatic polyp segmentation using convolution neural networks | [Paper](https://arxiv.org/pdf/2004.10792.pdf)/Code
2017 | **ISOP** | Fully convolutional neural networks for polyp segmentation in colonoscopy | [Paper](https://discovery.ucl.ac.uk/id/eprint/1540136/7/Rosa%20Brandao_101340F.pdf)/Code
2017 | **SPMB** | Superpixel based segmentation and classification of polyps in wireless capsule endoscopy | [Paper](https://arxiv.org/pdf/1710.07390.pdf)/Code
2016 | **IEEE TMI** | Convolutional Neural Networks for Medical Image Analysis: Full Training or Fine Tuning? | [Paper](https://arxiv.org/pdf/1706.00712.pdf)/Code
2016 | **ComNet** | Advanced Algorithm for Polyp Detection Using Depth Segmentation in Colon Endoscopy | [Paper*](https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7824010)/Code
2015 | **MICCAI** | U-net: Convolutional networks for biomedical image segmentation | [Paper](https://link.springer.com/content/pdf/10.1007/978-3-319-24574-4_28.pdf)/[Code](https://lmb.informatik.uni-freiburg.de/resources/opensource/unet/)
2015 | **CMIG** | WM-DOVA Maps for Accurate Polyp Highlighting in Colonoscopy: Validation vs. Saliency Maps from Physicians | [Paper](http://158.109.8.37/files/BSF2015.pdf)/[Code](https://polyp.grand-challenge.org/CVCClinicDB/)
2014 | **IJPRAI** | A complete system for candidate polyps detection in virtual colonoscopy | [Paper](https://arxiv.org/pdf/1209.6525.pdf)/Code
2014 | **IEEE TMI** | Automated polyp detection in colon capsule endoscopy | [Paper](https://arxiv.org/pdf/1305.1912.pdf)/Code
2012 | **PR** | Towards Automatic Polyp Detection with a Polyp Appearance Model | [Paper](http://refbase.cvc.uab.es/files/BSV2012a.pdf)/Code
2010 | **IEEE ICASSP** | Polyp detection in Wireless Capsule Endoscopy videos based on image segmentation and geometric feature | [Paper*](https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5495103)/Code
[Back to top](#1-preview)
## 2.2 Video Polyp Segmentation
**Yr.** | **Pub.** | **Title** | **Links**
:-: | :-: | :- | :-:
2022 | **MICCAI** | Semi-Supervised Spatial Temporal Attention Network for Video Polyp Segmentation | [Paper](https://link.springer.com/chapter/10.1007/978-3-031-16440-8_44)/Code
2022 | **AAAI** | TCCNet: Temporally Consistent Context-Free Network for Semi-supervised Video Polyp Segmentation | [Paper](https://www.ijcai.org/proceedings/2022/0155.pdf)/Code
2022 | **MIR** | :fire: Video Polyp Segmentation: A Deep Learning Perspective | [Paper](https://arxiv.org/pdf/2203.14291v3.pdf)/[Code](https://github.com/GewelsJI/VPS)
2021 | **MICCAI** | Progressively Normalized Self-Attention Network for Video Polyp Segmentation | [Paper](https://arxiv.org/pdf/2105.08468.pdf)/[Code](https://github.com/GewelsJI/PNS-Net)
2020 | **MICCAI** | Endoscopic Polyp Segmentation Using a Hybrid 2D/3D CNN | [Paper](https://discovery.ucl.ac.uk/id/eprint/10114066/1/Endoscopic%20polyp%20segmentation%20using%20a%20hybrid%202D-3D%20CNN.pdf)/Code
[Back to top](#1-preview)
## 2.3 Image Polyp Detection
**Yr.** | **Pub.** | **Title** | **Links**
:-: | :-: | :- | :-:
2022 | **JSJU** | Improving Colonoscopy Polyp Detection Rate Using Semi-Supervised Learning | [Paper](https://link.springer.com/content/pdf/10.1007/s12204-022-2519-1.pdf?pdf=inline%20link)/Code
2022 | **IJCARS** | Positive-gradient-weighted object activation mapping: visual explanation of object detector towards precise colorectal-polyp localisation | [Paper](https://link.springer.com/content/pdf/10.1007/s11548-022-02696-y.pdf)/Code
2022 | **arXiv** | Colonoscopy polyp detection with massive endoscopic images | [Paper](https://arxiv.org/pdf/2202.08730.pdf)/Code
2021 | **arXiv** | Detecting, Localising and Classifying Polyps from Colonoscopy Videos using Deep Learning | [Paper](https://arxiv.org/pdf/2101.03285.pdf)/Code
2020 | **Scientific Data** | HyperKvasir, a comprehensive multi-class image and video dataset for gastrointestinal endoscopy | [Paper](https://www.nature.com/articles/s41597-020-00622-y.pdf)/[Project](https://datasets.simula.no/hyper-kvasir/)
2020 | **Scientific reports** | Real-time detection of colon polyps during colonoscopy using deep learning: systematic validation with four independent datasets | [Paper](https://www.nature.com/articles/s41598-020-65387-1)/Code
2020 | **IEEE ISBI** | Reduce false-positive rate by active learning for automatic polyp detection in colonoscopy videos | [Paper](https://www.researchgate.net/profile/Zhe-Guo-12/publication/322563091_Automatic_polyp_recognition_from_colonoscopy_images_based_on_bag_of_visual_words/links/5f9b60a7299bf1b53e512f47/Automatic-polyp-recognition-from-colonoscopy-images-based-on-bag-of-visual-words.pdf)/Code
2020 | **TransAI** | Artifact Detection in Endoscopic Video with Deep Convolutional Neural Networks | [Paper*](https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9253131)/Code
2019 | **IEEE Access** | Colonic Polyp Detection in Endoscopic Videos With Single Shot Detection Based Deep Convolutional Neural Network | [Paper](https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=8731913)/Code
2019 | **ICTAI** | An Efficient Spatial-Temporal Polyp Detection Framework for Colonoscopy Video | [Paper*](https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8995313)/Code
2018 | **arXiv** | Y-net: A deep convolutional neural network for polyp detection | [Paper](https://arxiv.org/pdf/1806.01907.pdf)/Code
2017 | **IEEE TMI** | Comparative validation of polyp detection methods in video colonoscopy: results from the MICCAI 2015 endoscopic vision challenge | [Paper](http://clok.uclan.ac.uk/17023/2/17023%20Final%20Version.pdf)/Code
2015 | **IEEE TMI** | Automated Polyp Detection in Colonoscopy Videos Using Shape and Context Information | [Paper](https://sci-hub.se/10.1109/tmi.2015.2487997)/Code
2009 | **Bildverarbeitung fur die Medizin** | Texturebased polyp detection in colonoscopy | [Paper](http://ftp.informatik.rwth-aachen.de/Publications/CEUR-WS/Vol-446/p346.pdf)/Code
2009 | **Proc. SPIE** | A comparison of blood vessel features and local binary patterns for colorectal polyp classification | [Paper](https://www.lfb.rwth-aachen.de/files/publications/2009/GRO09a.pdf)/Code
2007 | **IEEE ICIP** | Polyp detection in colonoscopy video using elliptical shape feature | [Paper](https://projet.liris.cnrs.fr/imagine/pub/proceedings/ICIP-2007/pdfs/0200465.pdf)/Code
[Back to top](#1-preview)
## 2.4 Video Polyp Detection
**Yr.** | **Pub.** | **Title** | **Links**
:-: | :-: | :- | :-:
2023 | **MICCAI** | Self-Supervised Polyp Re-Identification in Colonoscopy | [Paper](https://arxiv.org/pdf/2306.08591.pdf)/[Code]()
2023 | **MICCAI** | YONA: You Only Need One Adjacent Reference-frame for Accurate and Fast Video Polyp Detection | [Paper](https://arxiv.org/pdf/2306.03686.pdf)/[Code]()
2021 | **BSPC** | Real-time automatic polyp detection in colonoscopy using feature enhancement module and spatiotemporal similarity correlation unit | [Paper](https://sci-hub.se/10.1016/j.bspc.2021.102503)/Code
2021 | **EIO** | Real-time deep learning-based colorectal polyp localization on clinical video footage achievable with a wide array of hardware configurations | [Paper](https://www.thieme-connect.com/products/ejournals/pdf/10.1055/a-1388-6735.pdf)/Code
2021 | **MICCAI** | Multi-frame Collaboration for Effective Endoscopic Video Polyp Detection via Spatial-Temporal Feature Transformation | [Paper](https://arxiv.org/pdf/2107.03609.pdf)/[Code](https://github.com/lingyunwu14/STFT)
2020 | **IEEE ISBI** | Polyp detection in colonoscopy videos by bootstrapping via temporal consistency | [Paper](https://sci-hub.se/downloads/2020-06-30//41/ma2020.pdf?rand=61ab92dda2f6b?download=true)/Code
2020 | **JBHI** | Improving Automatic Polyp Detection Using CNN by Exploiting Temporal Dependency in Colonoscopy Video | [Paper](https://ntnuopen.ntnu.no/ntnu-xmlui/bitstream/handle/11250/2723541/Improving+Automatic+Polyp+Detection+Using+CNN.pdf?sequence=2)/Code
2020 | **NPJ Digital Medicine** | AI-doscopist: a real-time deep-learning-based algorithm for localising polyps in colonoscopy videos with edge computing devices | [Paper](https://www.nature.com/articles/s41746-020-0281-z.pdf)/Code
2020 | **MICCAI** | Asynchronous in Parallel Detection and Tracking (AIPDT): Real-Time Robust Polyp Detection | [Paper*](https://link.springer.com/content/pdf/10.1007%2F978-3-030-59716-0.pdf)/Code
2019 | **IEEE ISBI** | POLYP TRACKING IN VIDEO COLONOSCOPY USING OPTICAL FLOW WITH AN ON-THE-FLY TRAINED CNN | [Paper*](https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8759180)/Code
2017 | **JBHI** | Integrating Online and Offline Three-Dimensional Deep Learning for Automated Polyp Detection in Colonoscopy Videos | [Paper](http://www.cse.cuhk.edu.hk/~qdou/papers/2017/%5B2017%5D%5BJBHI%5DIntegrating%20online%20and%20offline%20three%20dimensional%20deep%20learning%20for%20automated%20polyp%20detection%20in%20colonoscopy%20videos.pdf)/Code
2015 | **IPMI** | A Comprehensive Computer-Aided Polyp Detection System for Colonoscopy Videos | [Paper](https://link.springer.com/content/pdf/10.1007/978-3-319-19992-4_25.pdf?pdf=inline%20link)/Code
2015 | **IEEE ISBI** | Automatic polyp detection in colonoscopy videos using an ensemble of convolutional neural networks | [Paper](https://www.researchgate.net/profile/Nima-Tajbakhsh/publication/283464973_Automatic_polyp_detection_in_colonoscopy_videos_using_an_ensemble_of_convolutional_neural_networks/links/5718b4a708aed43f63221b27/Automatic-polyp-detection-in-colonoscopy-videos-using-an-ensemble-of-convolutional-neural-networks.pdf)/Code
[Back to top](#1-preview)
## 2.5 Image Polyp Classification
**Yr.** | **Pub.** | **Title** | **Links**
:-: | :-: | :- | :-:
2022 | **MICCAI** | FFCNet: Fourier Transform-Based Frequency Learning and Complex Convolutional Network for Colon Disease Classification | [Paper](https://link.springer.com/chapter/10.1007/978-3-031-16437-8_8)/[Code](https://github.com/soleilssss/FFCNet)
2022 | **MICCAI** | Toward Clinically Assisted Colorectal Polyp Recognition via Structured Cross-Modal Representation Consistency | [Paper](https://link.springer.com/content/pdf/10.1007/978-3-031-16437-8_14.pdf)/[Code](https://github.com/WeijieMax/CPC-Trans)
2020 | **IEEE ISBI** | Photoshopping Colonoscopy Video Frames | [Paper](https://arxiv.org/pdf/1910.10345v1.pdf)/Code
2020 | **MICCAI** | Few-Shot Anomaly Detection for Polyp Frames from Colonoscopy | [Paper](https://arxiv.org/pdf/2006.14811.pdf)/[Code](https://github.com/tianyu0207/FSAD-Net%20)
2020 | **MICCAI** | Two-Stream Deep Feature Modelling for Automated Video Endoscopy Data Analysis | [Paper](https://arxiv.org/pdf/2007.05914.pdf)/Code
2014 | **JICARS** | Towards embedded detection of polyps in WCE images for early diagnosis of colorectal cancer | [Paper](https://hal.archives-ouvertes.fr/hal-00843459/document)/[Code](https://polyp.grand-challenge.org/EtisLarib/)
[Back to top](#1-preview)
## 2.6 Video Polyp Classification
**Yr.** | **Pub.** | **Title** | **Links**
:-: | :-: | :- | :-:
2022 | **MICCAI** | Contrastive Transformer-based Multiple Instance Learning for Weakly Supervised Polyp Frame Detection | [Paper](https://link.springer.com/content/pdf/10.1007/978-3-031-16437-8_9.pdf)/[Code](https://github.com/tianyu0207/weakly-polyp)
[Back to top](#1-preview)
## 2.7 Colonoscopy Depth Estimation
**Yr.** | **Pub.** | **Title** | **Links**
:-: | :-: | :- | :-:
2022 | **arXiv** | Task-Aware Active Learning for Endoscopic Image Analysis | [Paper](https://arxiv.org/pdf/2204.03440.pdf)/[Code](https://github.com/thetna/endo-active-learn)
2019 | **IJCARS** | Implicit domain adaptation with conditional generative adversarial networks for depth prediction in endoscopy | [Paper](https://link.springer.com/content/pdf/10.1007/s11548-019-01962-w.pdf?pdf=button%20sticky)/Code/[Project](http://cmic.cs.ucl.ac.uk/ColonoscopyDepth)
[Back to top](#1-preview)
## 2.8 Colonoscopy Deficient Coverage Detection
**Yr.** | **Pub.** | **Title** | **Links**
:-: | :-: | :- | :-:
2020 | **IEEE TMI** | Detecting Deficient Coverage in Colonoscopies | [Paper](https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9097918)/Code
[Back to top](#1-preview)
## 2.9 Colon Polyp Image Synthesis
**Yr.** | **Pub.** | **Title** | **Links**
:-: | :-: | :- | :-:
2018 | **IEEE Access** | Abnormal Colon Polyp Image Synthesis Using Conditional Adversarial Networks for Improved Detection Performance | [Paper](https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=8478237)/Code
[Back to top](#1-preview)
# 3. Colonoscopy Resources
## 3.1 Image Segmentation Datasets
**Yr.** | **Pub.** | **Title** | **Links**
:-: | :-: | :- | :-:
2022 | **arXiv** | ERS: a novel comprehensive endoscopy image dataset for machine learning, compliant with the MST 3.0 specification | [Paper](https://arxiv.org/abs/2201.08746)/[Project](https://cvlab.eti.pg.gda.pl/publications/endoscopy-dataset)
2022 | **arXiv** | Synthetic data for unsupervised polyp segmentation | [Paper](https://arxiv.org/pdf/2202.08680.pdf)/[Code](https://github.com/enric1994/synth-colon)/[Project](https://enric1994.github.io/synth-colon)
2020 | **ICMM** | **Kvasir-SEG** - Kvasir-seg: A segmented polyp dataset | [Paper](https://arxiv.org/pdf/1911.07069.pdf)/[Code](https://datasets.simula.no/kvasir-seg/)
2017 | **JHE** | **CVC-EndoSceneStill** - A Benchmark for Endoluminal Scene Segmentation of Colonoscopy Images | [Paper](https://downloads.hindawi.com/journals/jhe/2017/4037190.pdf)/[Project](http://www.cvc.uab.es/CVC-Colon/index.php/databases/cvc-endoscenestill/)
2015 | **CMIG** | **CVC-ClinicDB/CVC-612** - WM-DOVA Maps for Accurate Polyp Highlighting in Colonoscopy: Validation vs. Saliency Maps from Physicians | [Paper](http://158.109.8.37/files/BSF2015.pdf)/[Project](https://polyp.grand-challenge.org/CVCClinicDB/)
2014 | **JICARS** | **ETIS-Larib Polyp DB** - Towards embedded detection of polyps in WCE images for early diagnosis of colorectal cancer | [Paper](https://hal.archives-ouvertes.fr/hal-00843459/document)/[Project](https://polyp.grand-challenge.org/EtisLarib/)
2012 | **PR** | **CVC-ColonDB/CVC-300** - Towards Automatic Polyp Detection with a Polyp Appearance Model | [Paper](http://refbase.cvc.uab.es/files/BSV2012a.pdf)/[Project](http://mv.cvc.uab.es/projects/colon-qa/cvccolondb)
_ | _ | **PICCOLO** - PICCOLO RGB/NBI (WIDEFIELD) IMAGE COLLECTION | [Project](https://www.biobancovasco.org/en/Sample-and-data-catalog/Databases/PD178-PICCOLO-EN.html)
[Back to top](#1-preview)
## 3.2 Video Segmentation Datasets
**Yr.** | **Pub.** | **Title** | **Links**
:-: | :-: | :- | :-:
2017 | **EIO** | **KID Project** - KID Project: an internet-based digital video atlas of capsule endoscopy for research purposes | [Paper](https://www.thieme-connect.de/products/ejournals/pdf/10.1055/s-0043-105488.pdf)/[Project](https://mdss.uth.gr/datasets/endoscopy/kid/)
2015 | **IEEE TMI** | **ASU-Mayo** - Automated Polyp Detection in Colonoscopy Videos Using Shape and Context Information | [Paper](https://sci-hub.se/10.1109/tmi.2015.2487997)/[Project](https://polyp.grand-challenge.org/AsuMayo/)
[Back to top](#1-preview)
## 3.3 Video Detection Datasets
**Yr.** | **Pub.** | **Title** | **Links**
:-: | :-: | :- | :-:
2021 | **GE** | **SUN Dataset** - Development of a computer-aided detection system for colonoscopy and a publicly accessible large colonoscopy video database (with video) | [Paper](https://pdf.sciencedirectassets.com/273305/1-s2.0-S0016510721X0003X/1-s2.0-S0016510720346551/main.pdf?X-Amz-Security-Token=IQoJb3JpZ2luX2VjEKr%2F%2F%2F%2F%2F%2F%2F%2F%2F%2FwEaCXVzLWVhc3QtMSJHMEUCIQCoLFri%2FY2qg1dLegjpE85SraDAQgXni4AstwVHir31FQIgSZt7d3LRM%2FDWZnrG2ob5NXTOCC6qrgtukFoyETMmG60qgwQIg%2F%2F%2F%2F%2F%2F%2F%2F%2F%2F%2FARAEGgwwNTkwMDM1NDY4NjUiDB6thtIGUd1PJRcyGSrXA7YEHw4pAwRRjdUrDIBxA5b6lw7vwYqXu%2FkyL8wa5THz4ls%2BQ79EWG1w9zl7j9F6A9bkKlVvKCAb0oi3e03KthHn%2B8g0l4OC0qix4pb0UUwreyZgOjArd70QgeuNuGUMxagQQ0SWaQUG%2FO%2B24Zr7sqoJjCFuNyulHGwblX4JXXI9rhGeb2yWr%2FKRTmbwiCKzuerSPMtbyJGK72cZ5qWuriDfQfUoNqKp49hRkitn7ZzSrz0wayxDzK6PKgPXLzx60HsPBgz%2BcPDKsLlEKdrtnOHcpTzINtfACgeTkvm8QP5WQq9SQO4PNfOgWKMEBxkkXqNaXRlWCriDE8ikkIxTS1wg1bBzX6bbq2VXPQ1HWzoCozUUBpla1%2FNddRj3cOdWNPV1CMDZKivYiFQGuB5ARoL7ijrhNH0igSNRe2WKoerxDKdKfOVmaRm9TYwuqVN6jS%2B1nS%2Bd2yY090PHBWsBHK0ZC2ACs2gHTJdafVkDObbFKhyzU3%2B6Q3rVKCjC6Rw4sJnNz0xsDPfGKVZ%2Ffhh4QAzGdJi18NBSmADUbEEXgV8gYg6HgQvblxNqFcwrZqmCab0QYWvg6q0%2FqyJhEYcmVWEdQJr9wVCWHGNSe2%2BPFfKR51sURtmWBTCX17WNBjqlAXvCE4xPsYowWXOcK%2BWOREDfMffE6zUdWetTgKWtCjFFzqvbe%2BaeKWuwEHXQl9BuG1rERZT9fY9aEEEVE2q6W0cDvbkVuFnmXxmhpxYw6qlJisddMBTNDWrLB8llUO0sASdIj0uz5uKqE1zqL%2FLVS1CRv1fbYzdXwEyRqrHux3g%2BArKgJ5uq92X9jznB8E0mLTrlwVI7OYjZfLdkwIXvry008gFE9g%3D%3D&X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Date=20211206T024723Z&X-Amz-SignedHeaders=host&X-Amz-Expires=300&X-Amz-Credential=ASIAQ3PHCVTYQITTHIVX%2F20211206%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Signature=9dff7d4ff54b70360044a95447978a4d661455c0e978adb1d72f44a240f04b2c&hash=96a9a5b1c7f87db2e806041f936f9262c2f0f8a853107bf2f6ef69c9e1d1db35&host=68042c943591013ac2b2430a89b270f6af2c76d8dfd086a07176afe7c76c2c61&pii=S0016510720346551&tid=spdf-21d9a413-2601-4aaf-8815-4d4142eeaf04&sid=ec558a9b400130484b2a9f1-fbc4af95d4adgxrqa&type=client)/[Project](http://amed8k.sundatabase.org)
[Back to top](#1-preview)
## 3.4 Video Classification Datasets
**Yr.** | **Pub.** | **Title** | **Links**
:-: | :-: | :- | :-:
2020 | **Scientific Data** | **HyperKvasir** - HyperKvasir, a comprehensive multi-class image and video dataset for gastrointestinal endoscopy | [Paper](https://www.nature.com/articles/s41597-020-00622-y.pdf)/[Project](https://datasets.simula.no/hyper-kvasir/)
2017 | **ACM MSC** | **Kvasir** - KVASIR: A Multi-Class Image Dataset for Computer Aided Gastrointestinal Disease Detection | [Paper](https://dl.acm.org/doi/pdf/10.1145/3083187.3083212)/[Project](https://datasets.simula.no/kvasir/)
2017 | **IEEE TMI** | **Colonoscopic Dataset** - Computer-Aided Classification of Gastrointestinal Lesions in Regular Colonoscopy | [Paper](https://hal.archives-ouvertes.fr/hal-01291797/document)/[Project](http://www.depeca.uah.es/colonoscopy_dataset/)
[Back to top](#1-preview)
## 3.5 Colonoscopy Depth Datasets
**Yr.** | **Pub.** | **Title** | **Links**
:-: | :-: | :- | :-:
2020 | **IEEE TMI** | Detecting Deficient Coverage in Colonoscopies | [Paper](https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9097918)/Code/[Project](https://dl.google.com/datasets/CC20/Google-CC20-dataset.tar.gz)
2019 | **IJCARS** | Implicit domain adaptation with conditional generative adversarial networks for depth prediction in endoscopy | [Paper](https://link.springer.com/content/pdf/10.1007/s11548-019-01962-w.pdf)/[Project](http://cmic.cs.ucl.ac.uk/ColonoscopyDepth/)
[Back to top](#1-preview)
# 4. Useful Resources
## 4.1 Colonoscopy Related
- [Deep Learning for Colonoscopy](https://github.com/GewelsJI/deep-learning-colonoscopy)
## 4.2 AI Conference Deadlines
- [Acceptance Rate for AI Conferences](https://github.com/lixin4ever/Conference-Acceptance-Rate)
- [AI Conference Deadlines](https://aideadlin.es/?sub=ML,CV,NLP,RO,SP,DM)
[Back to top](#1-preview)
================================================
FILE: docs/DATA_DESCRIPTION.md
================================================
# The Descriptions of SUN-SEG Dataset
<p align="center">
<img src="../assets/video_v2-min.gif"/> <br />
</p>
We first introduce a high-quality per-frame annotated VPS dataset, named SUN-SEG, which includes 158,690 frames from the famous SUN dataset. We extend the labels with diverse types, i.e., object mask, boundary, scribble, polygon, and visual attribute. We also introduce the pathological information from the original [SUN dataset](http://sundatabase.org/), including pathological classification labels, location information, and shape information.
Notably, the origin SUN dataset has 113 colonoscopy videos, including 100 positive cases with 49, 136 polyp frames and 13 negative cases with 109, 554 non-polyp frames. We manually trim them into 378 positive and 728 negative short clips, meanwhile maintaining their intrinsic consecutive relationship. Such data pre-processing ensures each clip has around 3~11s duration at a real-time frame rate (i.e., 30 fps), which promotes the fault-tolerant margin for various algorithms and devices. To this end, the re-organized SUN-SEG contains 1, 106 short video clips with 158, 690 video frames totally, offering a solid foundation to build a representative benchmark.
As such, it yields the final version of our SUN-SEG dataset, which includes 49,136 polyp frames (i.e., positive part) and 109,554 non-polyp frames (i.e., negative part) taken from different 285 and 728 colonoscopy videos clips, as well as the corresponding annotations. The following sections will provide details about our SUN-SEG point-by-point.
- [The Descriptions of SUN-SEG Dataset](#the-descriptions-of-sun-seg-dataset)
- [File Tree Organization](#file-tree-organization)
- [Dataset Statistics](#dataset-statistics)
- [Positive Part](#positive-part)
- [Negative Part](#negative-part)
- [Label Description](#label-description)
- [Label-I: Category Classification Annotation](#label-i-category-classification-annotation)
- [Label-II: Object Mask](#label-ii-object-mask)
- [Label-III: Bounding Box](#label-iii-bounding-box)
- [Label-IV: Boundary](#label-iv-boundary)
- [Label-V: Two Weak Labels (Scribble & Polygon)](#label-v-two-weak-labels-scribble--polygon)
- [Label-VI: Attributes Description](#label-vi-attributes-description)
- [Rejected Labels](#rejected-labels)
- [Citations](#citations)
- [Reference](#reference)
# File Tree Organization
The `Frame` folder contains the frames and the rest folders contain the corresponding ground truth. As for the `bbox_annotation.json` and `classfication.txt` text files, we follow the same format as COCO and ImageNet for generality.
```
├──data
├──SUN-SEG
├──TrainDataset
├──Frame # The images from SUN dataset
├──case1_1
├──image_name_00001.jpg
|...
├──case1_3
|...
├──GT # Object-level segmentation mask
├──case1_1
├──image_name_00001.png
|...
├──case1_3
|...
├──Edge # Weak label with edge
|...
├──Scribble # Weak label with scribble
|...
├──Polygon # Weak label with Polygon
|...
├──Classification # Category classification annotation
├──classification.txt
├──Detection # Bounding box
├──bbox_annotation.json
├──TestEasyDataset
├──Seen
├──Frame
├──case2_3
|...
├──GT
├──case2_3
|...
|...
├──Unseen
├──Frame
├──case3_1
|...
├──GT
├──case3_1
|...
|...
├──TestHardDataset
├──Seen
├──Frame
├──case1_2
|...
├──GT
├──case1_2
|...
|...
├──Unseen
├──Frame
├──case10_1
|...
├──GT
├──case10_1
|...
|...
```
# Dataset Statistics
Figure 1 (left) shows the statistic distributions for pathological patterns excluding non-polyp (NP). We find that well-differentiated or low-grade adenoma is dominated but is difficult to locate due to the low-intensity contrast between the lesion and mucosal surface. Figure 1 (right) shows the multi-dependencies among pathological patterns, shape, and location of colon polyp.
<p align="center">
<img src="../assets/DataStatistic.png"/> <br />
<em>
Figure 1: (Left) Distribution over pathological patterns. (Right) Multi-dependencies among pathological pattern, shape, and location.
</em>
</p>
## Positive Part
- The positive part of SUN-SEG has 285 video clips (30 fps), which has 49,136 frames.
- More details of each polyp video clips refer to [`INFO_POSITIVE_CASES.md`](https://github.com/GewelsJI/VPS/blob/main/docs/INFO_POSITIVE_CASES.md).
## Negative Part
- The negative part of SUN-SEG has 728 video clips (30 fps), which has 109,554 frames.
- More details of each non-polyp video clips refer to [`INFO_NEGATIVE_CASES.md`](https://github.com/GewelsJI/VPS/blob/main/docs/INFO_NEGATIVE_CASES.md)
# Label Description
## Label-I: Category Classification Annotation
<p align="center">
<img src="../assets/classification-min.png"/> <br />
</p>
Here are seven classes of pathological diagnosis:
- Low-grade adenoma (229 videos, 39834 frames)
- High-grade adenoma (26 videos, 4111 frames)
- Hyperplastic polyp (10 videos, 1644 frames)
- Traditional serrated adenoma (9 videos, 1627 frames)
- Sessile serrated lesion (8 videos, 1288 frames)
- Invasive carcinoma (3 videos, 632 frames)
- Non-Polyp (728 videos, 109,554 frames)
The annotation is in `./data/DATASET/classification.txt`.
In the text file, each row represents an image and its class of pathological diagnosis.
Here are an example:
image_dir_00001.jpg low_grade_adenoma
image_dir_00002.jpg hyperplastic_polyp
image_dir_00003.jpg sessile_serrated_lesion
...
## Label-II: Object Mask
<p align="center">
<img src="../assets/video_v2-min.gif"/> <br />
</p>
In polyp-existing frames, each polyp is annotated with a segmentation mask as shown above.
The annotation is in `./data/DATASET/GT/`. Each image's name has a direct correspondence with the annotation file name.
For example, the segmentation mask for `image_dir_00001.jpg` is `image_dir_00001.png`.
## Label-III: Bounding Box
<p align="center">
<img src="../assets/bbox-min.gif"/> <br />
</p>
We present the bounding box annotation for each polyp-existing frame. In `./data/DATASET/bbox_annotation.json` file, we follow the same format as the COCO dataset. Here is an example of COCO-style annotation:
{
'info': {
'year': 2021,
'version': 'v1.0',
'description': 'SUN Colonoscopy Video Database. Hayato et al, 2020.',
'contributor': '',
'url': '',
'date_created': ''},
'images': [{
'id': 'case1_1-a2-image0001',
'width': 1158,
'height': 1008,
'case_name': 'case1_1' # case_name means the name of case in the folder.
'file_name': 'case_M_20181001100941_0U62372100109341_1_005_001-1_a2_ayy_image0001'}, # file_name is corresponding to the image name in the folder.
...],
'annotation': [{
'id': 'case1_1-a2-image0001',
'bbox': [72, 262, 68, 81]}, # Each element represnets the [min_x, min_y, width, height], where min_x and min_y are the upper-left coordinates of the bounding box.
...]
}
## Label-IV: Boundary
The annotations are stored in `./data/DATASET/Edge/`. Each image's name has a direct correspondence with the annotation file name.
<p align="center">
<img src="../assets/weak1-min.gif"/> <br />
</p>
## Label-V: Two Weak Labels (Scribble & Polygon)
The annotations are in `./data/DATASET/Scribble/`, and `./data/DATASET/Polygon/`, respectively. Each image's name has a direct correspondence with the annotation file name.
<p align="center">
<img src="../assets/weak2-min.gif"/> <br />
</p>
## Label-VI: Attributes Description
Next, we provide the complete attributes for our SUN-SEG dataset.
- **Pathological Patterns**
| ID | Name | Description |
| ---- | ---------------------------- |-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| LA | Low-grade adenoma | The polyp with low-grade dysplasia often shows nuclear changes, such as palisading and darkening of the nucleus. |
| HA | High-grade adenoma | The polyp with high-grade dysplasia, which has more severe cellular and nuclear changes. |
| HP | Hyperplastic polyp | The polyp has small vessels or sparse networks, with unrecognizable patterns and is lighter than or similar to the surroundings. |
| TSA | Traditional serrated adenoma | A neoplastic polyp characterised by eosinophilic cells, ectopic crypt formations and slit-like epithelial serrations. |
| SSL | Sessile serrated lesion | A neoplastic polyp characterised by serrated architectural features and lack of cytological dysplasia. |
| IC | Invasive cancer (T1b) | Its colour is darker than the surroundings, brownish, sometimes with lighter patches. The vessel of areas with interrupted or absent vessels. The surface is amorphous with no surface pattern. |
| SI | Surgical Instruments | The endoscopic surgical procedures involve the positioning of instruments, such as snares, forceps, knives and electrodes. |
- **Shape**
> We follow the Narrow Band Imaging International Colorectal Endoscopic (NICE) classification criteria. It uses staining, vascular patterns, and surface patterns to distinguish between hyperplastic and adenomatous colon polyps. More details refer to [link-1](https://www.endoscopy-campus.com/en/classifications/polyp-classification-nice/) and [link-2](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5369434/)
<p align="center">
<img src="../assets/Shape.png"/> <br />
</p>
| ID | Name | Description |
| ---- | ---------------------------- |-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Ip | Pedunculated | Base is more narrow than the top of the lesion. |
| Isp | Subpedunculated | Intermediate and broad-based. Same management as (0-Is) sessile polyps. |
| Is | Sessile | Base and top of the lesion have the same diameter. |
| IIa | Slightly elevated | Lesion is slightly higher than adjacent mucosa. |
- **Location**
| ID | Name | Description |
| ---- | ---------------------------- |-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| C | Cecum | Lesion is located in Cecum. |
| A | Ascending colon | Lesion is located in Ascending colon. |
| T | Transverse colon | Lesion is located in the Transverse colon. |
| D | Descending colon | Lesion is located in Descending colon. |
| S | Sigmoid colon | Lesion is located in the Sigmoid colon. |
| R | Rectum | Lesion is located in Rectum. |
- **Visual Attributes**
| ID | Name | Description |
| ---- | ---------------------------- |--------------------------------------------------------------------------------------------------------------------------------------------------|
| IB | Indefinable Boundaries | The foreground and background areas around the object have a similar colour. |
| HO | Heterogeneous Object | Object regions have distinct colours. |
| GH | Ghosting | Object has anomaly RGB-colored boundary due to fast-moving or insufficient refresh rate. |
| FM | Fast-Motion | The average per-frame object motion, computed as the Euclidean distance of polyp centroids between consecutive frames, is larger than 20 pixels. |
| SO | Small-Object | The average ratio between the object size and the image area is smaller than 0.05. |
| LO | Large-Object | The average ratio between the object bounding-box area and the image area is larger than $t_{lr}$ = 0.15. |
| OCC | Occlusion | Object becomes partially or fully occluded. |
| OV | Out-of-view | Object is partially clipped by the image boundaries. |
| SV | Scale-Variation | The average area ratio among any pair of bounding boxes enclosing the target object is smaller than $0.5$. |
# Rejected Labels
To support learning strategies such as multi-rater agreement modeling, we follow the same file organization and relaese the rejected labels from our labeling process.
We release 49,136+17,422 labels in total, you can download the annotation file from [OneDrive](https://anu365-my.sharepoint.com/:u:/g/personal/u7248002_anu_edu_au/EQJgEgN7RLZMgOxIEjHOIUMBakeE3BUU6grmCG-J0r0IBQ?e=tk1XDz) / [Baidu Drive](https://pan.baidu.com/s/14v5LB7QrrhFm3JSwmoA2bQ) (Password: inqb, Size: 120MB). The label with `IMAGE_NAME_*.png` naming format is the rejected label and vice versa.
There are 33,331 images with 1 label, 14,211 images with 2 labels, 1571 images with 3 labels, and 23 images with 4 labels.
# Citations
If you have found our work useful, please use the following reference to cite this project:
@article{ji2022vps,
title={Video Polyp Segmentation: A Deep Learning Perspective},
author={Ji, Ge-Peng and Xiao, Guobao and Chou, Yu-Cheng and Fan, Deng-Ping and Zhao, Kai and Chen, Geng and Fu, Huazhu and Van Gool, Luc},
journal={Machine Intelligence Research},
year={2022}
}
@inproceedings{ji2021pnsnet,
title={Progressively Normalized Self-Attention Network for Video Polyp Segmentation},
author={Ji, Ge-Peng and Chou, Yu-Cheng and Fan, Deng-Ping and Chen, Geng and Jha, Debesh and Fu, Huazhu and Shao, Ling},
booktitle={MICCAI},
pages={142--152},
year={2021}
}
@article{misawa2021development,
title={Development of a computer-aided detection system for colonoscopy and a publicly accessible large colonoscopy video database (with video)},
author={Misawa, Masashi and Kudo, Shin-ei and Mori, Yuichi and Hotta, Kinichi and Ohtsuka, Kazuo and Matsuda, Takahisa and Saito, Shoichi and Kudo, Toyoki and Baba, Toshiyuki and Ishida, Fumio and others},
journal={Gastrointestinal endoscopy},
volume={93},
number={4},
pages={960--967},
year={2021},
publisher={Elsevier}
}
## Reference
- SUN dataset: http://sundatabase.org
- COCO dataset: https://cocodataset.org
================================================
FILE: docs/DATA_PREPARATION.md
================================================
# Dataset Preparation
We introduce a high-quality per-frame annotated VPS dataset, named SUN-SEG, which includes 158,690 frames elected from the famous [SUN dataset](http://amed8k.sundatabase.org). Then, we extend the labels with diverse types, i.e., object mask, boundary, scribble, and polygon. If you want to get access to our whole dataset, you should follow the next three steps.
# Contents
- [Step-1: Request and Download](#step-1--request-and-download)
- [Step-2: Unzip SUN dataset](#step-2--unzip-sun-dataset)
- [Step-3: Re-organize the file structure](#step-3--re-organize-the-file-structure)
# Step-1: Request the raw SUN-database
The origin colonoscopy video frames in our SUN-SEG dataset are selected from [SUN dataset](http://amed8k.sundatabase.org), while we could not distribute the video data due to the strict license. So first, you guys need to request the original colonoscopy video frame from them. In this step, you should download the polyp samples of 100 cases and non-polyp samples of 13 cases from the links provided by the SUN dataset.
- **Request for video frames from SUN:** Please follow the instruction on [SUN dataset](http://amed8k.sundatabase.org) to request SUN-dataset and download the dataset by yourself. **Please use your educational email to apply for it and claim it without any commercial purpose.** Thank you for your understanding! Let me know (gepengai.ji@gmail.com) if you have any questions.
# ⭐ Step-2: Download our annotations
Then, you need to download the complete annotation provided in our SUN-SEG, which can be downloaded at the download link:
- (July 17, 2024) Updated link: [annotation.v2 - google drive](https://drive.google.com/file/d/1ytmhpg0YaW0XZBAEfSFkMhaI6jnGLSg3/view?usp=sharing)
# Step-3: Unzip SUN dataset
As for video frames in the SUN dataset, these are two groups of samples, which are divided into multiple compressed file formats as zip files. To decompress each zip file downloaded, please input the password provided by the origin authors of the SUN dataset (i.e., the same as the password that you used for login).
- **Unzip positive cases in SUN**
- create directory: `mkdir ./data/SUN-Positive/`
- unzip positive cases: `unzip -P sun_password -d ./SUN-Positive sundatabase_positive_part\*`, which will take up 11.5 + 9.7 GB of storage space. Please ensure your server has enough space to store them, otherwise, it will fail. Please replace the `sun_password` that you get from SUN's authors.
- check if correct: `find ./SUN-Positive -type f -name "*.jpg" | wc -l`, which should output 49,136 in your terminal.
- **Unzip negative cases in SUN (Optional)**
- create directory: `mkdir ./data/SUN-Negative/`
- unzip negative cases: `unzip -P sun_password -d ./SUN-Negative sundatabase_positive_part\*`, which will take up 11.6 + 10.7 + 11.5 + 10.5 GB of storage space. (This data partition is optional if we have no requirements to use them.)
- check if correct: `find ./SUN-Negative -type f -name "*.jpg" | wc -l`, which should output 109,554 in your terminal.
As for the annotations from our SUN-SGE, you are happy to execute:
- Unwarp it via `unzip SUN-SEG-Annotation-v2.zip`
- Put it at path `./data/SUN-SEG-Annotation/`
After preparing all the files, your file structure will be the same as below:
```
├──data
├──SUN-Positive
├──case1
├──image_name.jpg
|...
├──case2
|...
├──SUN-SEG-Annotation
├──TrainDataset
├──GT
├──case1_1
├──image_name.png
|...
├──Edge
|...
├──Scribble
|...
├──Polygon
|...
├──Classification
├──classification.txt
├──Detection
├──bbox_annotation.json
├──TestEasyDataset
├──Seen
├──GT
|...
├──Unseen
├──GT
|...
├──TestHardDataset
├──Seen
├──GT
|...
├──Unseen
├──GT
|...
```
You will notice that the file structure of images in `SUN-Positive` is different from the one of annotation in `SUN-SEG-Annotation`. To reconcile the file structure, you need to follow next step to finish the data preparation.
# Step-4: Re-organize the file structure
By running `sh ./utils/reorganize.sh`, the original file structure in SUN-dataset will be reorganised to the same as SUN-SEG for better length balance. Finally, the folder `Frame` which is originated from `SUN-Positive`, and `GT`, as long as other annotations folders, will share the same file structure as shown below:
```
├──data
├──SUN-SEG
├──TrainDataset
├──Frame # The images from the SUN dataset
├──case1_1
├──image_name_00001.jpg
|...
├──case1_3
|...
├──GT # Object-level segmentation mask
├──case1_1
├──image_name_00001.png
|...
├──case1_3
|...
├──Edge # Weak label with edge
|...
├──Scribble # Weak label with scribble
|...
├──Polygon # Weak label with Polygon
|...
├──Classification # Category classification annotation
├──classification.txt
├──Detection # Bounding box
├──bbox_annotation.json
├──TestEasyDataset
├──Seen
├──Frame
├──case2_3
|...
├──GT
├──case2_3
|...
|...
├──Unseen
├──Frame
├──case3_1
|...
├──GT
├──case3_1
|...
|...
├──TestHardDataset
├──Seen
├──Frame
├──case1_2
|...
├──GT
├──case1_2
|...
|...
├──Unseen
├──Frame
├──case10_1
|...
├──GT
├──case10_1
|...
|...
```
================================================
FILE: docs/INFO_NEGATIVE_CASES.md
================================================
# Information of Negative Cases in our SUN-SEG
| Clip-ID | Number of Frames | Duration (seconds) |
|----|------------------|-------------------------------|
| 1_1 | 150 | 5.00 |
| 1_2 | 150 | 5.00 |
| 1_3 | 150 | 5.00 |
| 1_4 | 150 | 5.00 |
| 1_5 | 150 | 5.00 |
| 1_6 | 150 | 5.00 |
| 1_7 | 150 | 5.00 |
| 1_8 | 150 | 5.00 |
| 1_9 | 150 | 5.00 |
| 1_10 | 150 | 5.00 |
| 1_11 | 150 | 5.00 |
| 1_12 | 150 | 5.00 |
| 1_13 | 150 | 5.00 |
| 1_14 | 150 | 5.00 |
| 1_15 | 150 | 5.00 |
| 1_16 | 150 | 5.00 |
| 1_17 | 150 | 5.00 |
| 1_18 | 150 | 5.00 |
| 1_19 | 150 | 5.00 |
| 1_20 | 150 | 5.00 |
| 1_21 | 150 | 5.00 |
| 1_22 | 150 | 5.00 |
| 1_23 | 150 | 5.00 |
| 1_24 | 150 | 5.00 |
| 1_25 | 150 | 5.00 |
| 1_26 | 150 | 5.00 |
| 1_27 | 150 | 5.00 |
| 1_28 | 150 | 5.00 |
| 1_29 | 150 | 5.00 |
| 1_30 | 150 | 5.00 |
| 1_31 | 150 | 5.00 |
| 1_32 | 150 | 5.00 |
| 1_33 | 150 | 5.00 |
| 1_34 | 150 | 5.00 |
| 1_35 | 150 | 5.00 |
| 1_36 | 150 | 5.00 |
| 1_37 | 150 | 5.00 |
| 1_38 | 150 | 5.00 |
| 1_39 | 150 | 5.00 |
| 1_40 | 150 | 5.00 |
| 1_41 | 150 | 5.00 |
| 1_42 | 150 | 5.00 |
| 1_43 | 150 | 5.00 |
| 1_44 | 150 | 5.00 |
| 1_45 | 150 | 5.00 |
| 1_46 | 150 | 5.00 |
| 1_47 | 150 | 5.00 |
| 1_48 | 150 | 5.00 |
| 1_49 | 150 | 5.00 |
| 1_50 | 150 | 5.00 |
| 1_51 | 150 | 5.00 |
| 1_52 | 150 | 5.00 |
| 1_53 | 150 | 5.00 |
| 1_54 | 150 | 5.00 |
| 1_55 | 150 | 5.00 |
| 1_56 | 150 | 5.00 |
| 1_57 | 150 | 5.00 |
| 1_58 | 150 | 5.00 |
| 1_59 | 150 | 5.00 |
| 1_60 | 150 | 5.00 |
| 1_61 | 150 | 5.00 |
| 1_62 | 150 | 5.00 |
| 1_63 | 150 | 5.00 |
| 1_64 | 150 | 5.00 |
| 1_65 | 150 | 5.00 |
| 1_66 | 211 | 7.03 |
| 2_1 | 150 | 5.00 |
| 2_2 | 150 | 5.00 |
| 2_3 | 150 | 5.00 |
| 2_4 | 150 | 5.00 |
| 2_5 | 150 | 5.00 |
| 2_6 | 150 | 5.00 |
| 2_7 | 150 | 5.00 |
| 2_8 | 150 | 5.00 |
| 2_9 | 150 | 5.00 |
| 2_10 | 150 | 5.00 |
| 2_11 | 150 | 5.00 |
| 2_12 | 150 | 5.00 |
| 2_13 | 150 | 5.00 |
| 2_14 | 150 | 5.00 |
| 2_15 | 150 | 5.00 |
| 2_16 | 150 | 5.00 |
| 2_17 | 150 | 5.00 |
| 2_18 | 150 | 5.00 |
| 2_19 | 150 | 5.00 |
| 2_20 | 150 | 5.00 |
| 2_21 | 150 | 5.00 |
| 2_22 | 150 | 5.00 |
| 2_23 | 150 | 5.00 |
| 2_24 | 150 | 5.00 |
| 2_25 | 150 | 5.00 |
| 2_26 | 150 | 5.00 |
| 2_27 | 150 | 5.00 |
| 2_28 | 150 | 5.00 |
| 2_29 | 150 | 5.00 |
| 2_30 | 150 | 5.00 |
| 2_31 | 150 | 5.00 |
| 2_32 | 150 | 5.00 |
| 2_33 | 150 | 5.00 |
| 2_34 | 150 | 5.00 |
| 2_35 | 150 | 5.00 |
| 2_36 | 150 | 5.00 |
| 2_37 | 150 | 5.00 |
| 2_38 | 150 | 5.00 |
| 2_39 | 150 | 5.00 |
| 2_40 | 150 | 5.00 |
| 2_41 | 150 | 5.00 |
| 2_42 | 150 | 5.00 |
| 2_43 | 150 | 5.00 |
| 2_44 | 150 | 5.00 |
| 2_45 | 150 | 5.00 |
| 2_46 | 150 | 5.00 |
| 2_47 | 150 | 5.00 |
| 2_48 | 150 | 5.00 |
| 2_49 | 150 | 5.00 |
| 2_50 | 150 | 5.00 |
| 2_51 | 150 | 5.00 |
| 2_52 | 150 | 5.00 |
| 2_53 | 150 | 5.00 |
| 2_54 | 150 | 5.00 |
| 2_55 | 150 | 5.00 |
| 2_56 | 150 | 5.00 |
| 2_57 | 150 | 5.00 |
| 2_58 | 150 | 5.00 |
| 2_59 | 150 | 5.00 |
| 2_60 | 150 | 5.00 |
| 2_61 | 150 | 5.00 |
| 2_62 | 150 | 5.00 |
| 2_63 | 150 | 5.00 |
| 2_64 | 150 | 5.00 |
| 2_65 | 150 | 5.00 |
| 2_66 | 150 | 5.00 |
| 2_67 | 173 | 5.77 |
| 3_1 | 150 | 5.00 |
| 3_2 | 150 | 5.00 |
| 3_3 | 150 | 5.00 |
| 3_4 | 150 | 5.00 |
| 3_5 | 150 | 5.00 |
| 3_6 | 150 | 5.00 |
| 3_7 | 150 | 5.00 |
| 3_8 | 150 | 5.00 |
| 3_9 | 150 | 5.00 |
| 3_10 | 150 | 5.00 |
| 3_11 | 150 | 5.00 |
| 3_12 | 150 | 5.00 |
| 3_13 | 150 | 5.00 |
| 3_14 | 150 | 5.00 |
| 3_15 | 150 | 5.00 |
| 3_16 | 150 | 5.00 |
| 3_17 | 150 | 5.00 |
| 3_18 | 150 | 5.00 |
| 3_19 | 150 | 5.00 |
| 3_20 | 150 | 5.00 |
| 3_21 | 150 | 5.00 |
| 3_22 | 150 | 5.00 |
| 3_23 | 150 | 5.00 |
| 3_24 | 150 | 5.00 |
| 3_25 | 150 | 5.00 |
| 3_26 | 150 | 5.00 |
| 3_27 | 150 | 5.00 |
| 3_28 | 150 | 5.00 |
| 3_29 | 150 | 5.00 |
| 3_30 | 150 | 5.00 |
| 3_31 | 150 | 5.00 |
| 3_32 | 150 | 5.00 |
| 3_33 | 150 | 5.00 |
| 3_34 | 150 | 5.00 |
| 3_35 | 150 | 5.00 |
| 3_36 | 150 | 5.00 |
| 3_37 | 150 | 5.00 |
| 3_38 | 150 | 5.00 |
| 3_39 | 150 | 5.00 |
| 3_40 | 150 | 5.00 |
| 3_41 | 150 | 5.00 |
| 3_42 | 150 | 5.00 |
| 3_43 | 150 | 5.00 |
| 3_44 | 150 | 5.00 |
| 3_45 | 150 | 5.00 |
| 3_46 | 150 | 5.00 |
| 3_47 | 150 | 5.00 |
| 3_48 | 102 | 3.40 |
| 4_1 | 150 | 5.00 |
| 4_2 | 150 | 5.00 |
| 4_3 | 150 | 5.00 |
| 4_4 | 150 | 5.00 |
| 4_5 | 150 | 5.00 |
| 4_6 | 150 | 5.00 |
| 4_7 | 150 | 5.00 |
| 4_8 | 150 | 5.00 |
| 4_9 | 150 | 5.00 |
| 4_10 | 150 | 5.00 |
| 4_11 | 150 | 5.00 |
| 4_12 | 150 | 5.00 |
| 4_13 | 150 | 5.00 |
| 4_14 | 150 | 5.00 |
| 4_15 | 150 | 5.00 |
| 4_16 | 150 | 5.00 |
| 4_17 | 150 | 5.00 |
| 4_18 | 150 | 5.00 |
| 4_19 | 150 | 5.00 |
| 4_20 | 150 | 5.00 |
| 4_21 | 150 | 5.00 |
| 4_22 | 150 | 5.00 |
| 4_23 | 150 | 5.00 |
| 4_24 | 150 | 5.00 |
| 4_25 | 150 | 5.00 |
| 4_26 | 150 | 5.00 |
| 4_27 | 150 | 5.00 |
| 4_28 | 150 | 5.00 |
| 4_29 | 150 | 5.00 |
| 4_30 | 150 | 5.00 |
| 4_31 | 150 | 5.00 |
| 4_32 | 150 | 5.00 |
| 4_33 | 150 | 5.00 |
| 4_34 | 150 | 5.00 |
| 4_35 | 150 | 5.00 |
| 4_36 | 150 | 5.00 |
| 4_37 | 150 | 5.00 |
| 4_38 | 150 | 5.00 |
| 4_39 | 150 | 5.00 |
| 4_40 | 150 | 5.00 |
| 4_41 | 150 | 5.00 |
| 4_42 | 150 | 5.00 |
| 4_43 | 150 | 5.00 |
| 4_44 | 150 | 5.00 |
| 4_45 | 150 | 5.00 |
| 4_46 | 150 | 5.00 |
| 4_47 | 150 | 5.00 |
| 4_48 | 150 | 5.00 |
| 4_49 | 150 | 5.00 |
| 4_50 | 150 | 5.00 |
| 4_51 | 150 | 5.00 |
| 4_52 | 150 | 5.00 |
| 4_53 | 150 | 5.00 |
| 4_54 | 150 | 5.00 |
| 4_55 | 150 | 5.00 |
| 4_56 | 150 | 5.00 |
| 4_57 | 150 | 5.00 |
| 4_58 | 150 | 5.00 |
| 4_59 | 150 | 5.00 |
| 4_60 | 150 | 5.00 |
| 4_61 | 150 | 5.00 |
| 4_62 | 150 | 5.00 |
| 4_63 | 150 | 5.00 |
| 4_64 | 150 | 5.00 |
| 4_65 | 150 | 5.00 |
| 4_66 | 150 | 5.00 |
| 4_67 | 150 | 5.00 |
| 4_68 | 150 | 5.00 |
| 4_69 | 150 | 5.00 |
| 4_70 | 150 | 5.00 |
| 4_71 | 150 | 5.00 |
| 4_72 | 150 | 5.00 |
| 4_73 | 150 | 5.00 |
| 4_74 | 150 | 5.00 |
| 4_75 | 150 | 5.00 |
| 4_76 | 150 | 5.00 |
| 4_77 | 150 | 5.00 |
| 4_78 | 150 | 5.00 |
| 4_79 | 150 | 5.00 |
| 4_80 | 150 | 5.00 |
| 4_81 | 150 | 5.00 |
| 4_82 | 150 | 5.00 |
| 4_83 | 150 | 5.00 |
| 4_84 | 150 | 5.00 |
| 4_85 | 150 | 5.00 |
| 4_86 | 150 | 5.00 |
| 4_87 | 150 | 5.00 |
| 4_88 | 150 | 5.00 |
| 4_89 | 150 | 5.00 |
| 4_90 | 150 | 5.00 |
| 4_91 | 150 | 5.00 |
| 4_92 | 150 | 5.00 |
| 4_93 | 150 | 5.00 |
| 4_94 | 150 | 5.00 |
| 4_95 | 150 | 5.00 |
| 4_96 | 150 | 5.00 |
| 4_97 | 235 | 7.83 |
| 5_1 | 150 | 5.00 |
| 5_2 | 150 | 5.00 |
| 5_3 | 150 | 5.00 |
| 5_4 | 150 | 5.00 |
| 5_5 | 150 | 5.00 |
| 5_6 | 150 | 5.00 |
| 5_7 | 150 | 5.00 |
| 5_8 | 150 | 5.00 |
| 5_9 | 150 | 5.00 |
| 5_10 | 150 | 5.00 |
| 5_11 | 150 | 5.00 |
| 5_12 | 150 | 5.00 |
| 5_13 | 150 | 5.00 |
| 5_14 | 150 | 5.00 |
| 5_15 | 150 | 5.00 |
| 5_16 | 150 | 5.00 |
| 5_17 | 150 | 5.00 |
| 5_18 | 150 | 5.00 |
| 5_19 | 150 | 5.00 |
| 5_20 | 150 | 5.00 |
| 5_21 | 150 | 5.00 |
| 5_22 | 150 | 5.00 |
| 5_23 | 150 | 5.00 |
| 5_24 | 150 | 5.00 |
| 5_25 | 150 | 5.00 |
| 5_26 | 150 | 5.00 |
| 5_27 | 150 | 5.00 |
| 5_28 | 150 | 5.00 |
| 5_29 | 150 | 5.00 |
| 5_30 | 150 | 5.00 |
| 5_31 | 150 | 5.00 |
| 5_32 | 150 | 5.00 |
| 5_33 | 150 | 5.00 |
| 5_34 | 150 | 5.00 |
| 5_35 | 150 | 5.00 |
| 5_36 | 150 | 5.00 |
| 5_37 | 150 | 5.00 |
| 5_38 | 150 | 5.00 |
| 5_39 | 150 | 5.00 |
| 5_40 | 150 | 5.00 |
| 5_41 | 150 | 5.00 |
| 5_42 | 150 | 5.00 |
| 5_43 | 150 | 5.00 |
| 5_44 | 150 | 5.00 |
| 5_45 | 150 | 5.00 |
| 5_46 | 150 | 5.00 |
| 5_47 | 150 | 5.00 |
| 5_48 | 150 | 5.00 |
| 5_49 | 150 | 5.00 |
| 5_50 | 150 | 5.00 |
| 5_51 | 150 | 5.00 |
| 5_52 | 150 | 5.00 |
| 5_53 | 116 | 3.87 |
| 6_1 | 150 | 5.00 |
| 6_2 | 150 | 5.00 |
| 6_3 | 150 | 5.00 |
| 6_4 | 150 | 5.00 |
| 6_5 | 150 | 5.00 |
| 6_6 | 150 | 5.00 |
| 6_7 | 150 | 5.00 |
| 6_8 | 150 | 5.00 |
| 6_9 | 150 | 5.00 |
| 6_10 | 150 | 5.00 |
| 6_11 | 150 | 5.00 |
| 6_12 | 150 | 5.00 |
| 6_13 | 150 | 5.00 |
| 6_14 | 150 | 5.00 |
| 6_15 | 150 | 5.00 |
| 6_16 | 150 | 5.00 |
| 6_17 | 150 | 5.00 |
| 6_18 | 150 | 5.00 |
| 6_19 | 150 | 5.00 |
| 6_20 | 150 | 5.00 |
| 6_21 | 150 | 5.00 |
| 6_22 | 150 | 5.00 |
| 6_23 | 150 | 5.00 |
| 6_24 | 150 | 5.00 |
| 6_25 | 150 | 5.00 |
| 6_26 | 150 | 5.00 |
| 6_27 | 150 | 5.00 |
| 6_28 | 150 | 5.00 |
| 6_29 | 150 | 5.00 |
| 6_30 | 150 | 5.00 |
| 6_31 | 150 | 5.00 |
| 6_32 | 150 | 5.00 |
| 6_33 | 150 | 5.00 |
| 6_34 | 150 | 5.00 |
| 6_35 | 150 | 5.00 |
| 6_36 | 150 | 5.00 |
| 6_37 | 150 | 5.00 |
| 6_38 | 150 | 5.00 |
| 6_39 | 150 | 5.00 |
| 6_40 | 150 | 5.00 |
| 6_41 | 150 | 5.00 |
| 6_42 | 150 | 5.00 |
| 6_43 | 150 | 5.00 |
| 6_44 | 150 | 5.00 |
| 6_45 | 150 | 5.00 |
| 6_46 | 150 | 5.00 |
| 6_47 | 150 | 5.00 |
| 6_48 | 150 | 5.00 |
| 6_49 | 150 | 5.00 |
| 6_50 | 150 | 5.00 |
| 6_51 | 150 | 5.00 |
| 6_52 | 150 | 5.00 |
| 6_53 | 150 | 5.00 |
| 6_54 | 150 | 5.00 |
| 6_55 | 150 | 5.00 |
| 6_56 | 150 | 5.00 |
| 6_57 | 150 | 5.00 |
| 6_58 | 150 | 5.00 |
| 6_59 | 150 | 5.00 |
| 6_60 | 150 | 5.00 |
| 6_61 | 150 | 5.00 |
| 6_62 | 150 | 5.00 |
| 6_63 | 150 | 5.00 |
| 6_64 | 150 | 5.00 |
| 6_65 | 150 | 5.00 |
| 6_66 | 150 | 5.00 |
| 6_67 | 150 | 5.00 |
| 6_68 | 150 | 5.00 |
| 6_69 | 150 | 5.00 |
| 6_70 | 150 | 5.00 |
| 6_71 | 150 | 5.00 |
| 6_72 | 150 | 5.00 |
| 6_73 | 150 | 5.00 |
| 6_74 | 150 | 5.00 |
| 6_75 | 150 | 5.00 |
| 6_76 | 150 | 5.00 |
| 6_77 | 150 | 5.00 |
| 6_78 | 150 | 5.00 |
| 6_79 | 150 | 5.00 |
| 6_80 | 150 | 5.00 |
| 6_81 | 150 | 5.00 |
| 6_82 | 150 | 5.00 |
| 6_83 | 150 | 5.00 |
| 6_84 | 150 | 5.00 |
| 6_85 | 150 | 5.00 |
| 6_86 | 150 | 5.00 |
| 6_87 | 150 | 5.00 |
| 6_88 | 150 | 5.00 |
| 6_89 | 150 | 5.00 |
| 6_90 | 150 | 5.00 |
| 6_91 | 150 | 5.00 |
| 6_92 | 150 | 5.00 |
| 6_93 | 150 | 5.00 |
| 6_94 | 150 | 5.00 |
| 6_95 | 150 | 5.00 |
| 6_96 | 150 | 5.00 |
| 6_97 | 150 | 5.00 |
| 6_98 | 150 | 5.00 |
| 6_99 | 150 | 5.00 |
| 6_100 | 150 | 5.00 |
| 6_101 | 150 | 5.00 |
| 6_102 | 150 | 5.00 |
| 6_103 | 150 | 5.00 |
| 6_104 | 150 | 5.00 |
| 6_105 | 150 | 5.00 |
| 6_106 | 150 | 5.00 |
| 6_107 | 150 | 5.00 |
| 6_108 | 150 | 5.00 |
| 6_109 | 150 | 5.00 |
| 6_110 | 150 | 5.00 |
| 6_111 | 150 | 5.00 |
| 6_112 | 150 | 5.00 |
| 6_113 | 150 | 5.00 |
| 6_114 | 96 | 3.20 |
| 7_1 | 150 | 5.00 |
| 7_2 | 150 | 5.00 |
| 7_3 | 150 | 5.00 |
| 7_4 | 150 | 5.00 |
| 7_5 | 150 | 5.00 |
| 7_6 | 150 | 5.00 |
| 7_7 | 150 | 5.00 |
| 7_8 | 150 | 5.00 |
| 7_9 | 150 | 5.00 |
| 7_10 | 150 | 5.00 |
| 7_11 | 150 | 5.00 |
| 7_12 | 150 | 5.00 |
| 7_13 | 150 | 5.00 |
| 7_14 | 150 | 5.00 |
| 7_15 | 150 | 5.00 |
| 7_16 | 150 | 5.00 |
| 7_17 | 150 | 5.00 |
| 7_18 | 150 | 5.00 |
| 7_19 | 150 | 5.00 |
| 7_20 | 150 | 5.00 |
| 7_21 | 150 | 5.00 |
| 7_22 | 150 | 5.00 |
| 7_23 | 150 | 5.00 |
| 7_24 | 150 | 5.00 |
| 7_25 | 150 | 5.00 |
| 7_26 | 150 | 5.00 |
| 7_27 | 150 | 5.00 |
| 7_28 | 150 | 5.00 |
| 7_29 | 150 | 5.00 |
| 7_30 | 150 | 5.00 |
| 7_31 | 150 | 5.00 |
| 7_32 | 150 | 5.00 |
| 7_33 | 150 | 5.00 |
| 7_34 | 150 | 5.00 |
| 7_35 | 150 | 5.00 |
| 7_36 | 150 | 5.00 |
| 7_37 | 236 | 7.87 |
| 8_1 | 150 | 5.00 |
| 8_2 | 150 | 5.00 |
| 8_3 | 150 | 5.00 |
| 8_4 | 150 | 5.00 |
| 8_5 | 150 | 5.00 |
| 8_6 | 150 | 5.00 |
| 8_7 | 150 | 5.00 |
| 8_8 | 150 | 5.00 |
| 8_9 | 150 | 5.00 |
| 8_10 | 150 | 5.00 |
| 8_11 | 150 | 5.00 |
| 8_12 | 150 | 5.00 |
| 8_13 | 150 | 5.00 |
| 8_14 | 150 | 5.00 |
| 8_15 | 150 | 5.00 |
| 8_16 | 150 | 5.00 |
| 8_17 | 168 | 5.60 |
| 9_1 | 150 | 5.00 |
| 9_2 | 150 | 5.00 |
| 9_3 | 150 | 5.00 |
| 9_4 | 150 | 5.00 |
| 9_5 | 150 | 5.00 |
| 9_6 | 150 | 5.00 |
| 9_7 | 150 | 5.00 |
| 9_8 | 150 | 5.00 |
| 9_9 | 150 | 5.00 |
| 9_10 | 150 | 5.00 |
| 9_11 | 150 | 5.00 |
| 9_12 | 150 | 5.00 |
| 9_13 | 150 | 5.00 |
| 9_14 | 150 | 5.00 |
| 9_15 | 150 | 5.00 |
| 9_16 | 150 | 5.00 |
| 9_17 | 150 | 5.00 |
| 9_18 | 150 | 5.00 |
| 9_19 | 150 | 5.00 |
| 9_20 | 150 | 5.00 |
| 9_21 | 150 | 5.00 |
| 9_22 | 150 | 5.00 |
| 9_23 | 150 | 5.00 |
| 9_24 | 150 | 5.00 |
| 9_25 | 150 | 5.00 |
| 9_26 | 150 | 5.00 |
| 9_27 | 150 | 5.00 |
| 9_28 | 150 | 5.00 |
| 9_29 | 150 | 5.00 |
| 9_30 | 150 | 5.00 |
| 9_31 | 150 | 5.00 |
| 9_32 | 150 | 5.00 |
| 9_33 | 150 | 5.00 |
| 9_34 | 150 | 5.00 |
| 9_35 | 150 | 5.00 |
| 9_36 | 150 | 5.00 |
| 9_37 | 150 | 5.00 |
| 9_38 | 150 | 5.00 |
| 9_39 | 150 | 5.00 |
| 9_40 | 150 | 5.00 |
| 9_41 | 150 | 5.00 |
| 9_42 | 150 | 5.00 |
| 9_43 | 150 | 5.00 |
| 9_44 | 150 | 5.00 |
| 9_45 | 150 | 5.00 |
| 9_46 | 150 | 5.00 |
| 9_47 | 150 | 5.00 |
| 9_48 | 150 | 5.00 |
| 9_49 | 150 | 5.00 |
| 9_50 | 150 | 5.00 |
| 9_51 | 150 | 5.00 |
| 9_52 | 150 | 5.00 |
| 9_53 | 150 | 5.00 |
| 9_54 | 150 | 5.00 |
| 9_55 | 150 | 5.00 |
| 9_56 | 150 | 5.00 |
| 9_57 | 150 | 5.00 |
| 9_58 | 150 | 5.00 |
| 9_59 | 150 | 5.00 |
| 9_60 | 150 | 5.00 |
| 9_61 | 150 | 5.00 |
| 9_62 | 150 | 5.00 |
| 9_63 | 222 | 7.40 |
| 10_1 | 150 | 5.00 |
| 10_2 | 150 | 5.00 |
| 10_3 | 150 | 5.00 |
| 10_4 | 150 | 5.00 |
| 10_5 | 150 | 5.00 |
| 10_6 | 150 | 5.00 |
| 10_7 | 150 | 5.00 |
| 10_8 | 150 | 5.00 |
| 10_9 | 150 | 5.00 |
| 10_10 | 150 | 5.00 |
| 10_11 | 150 | 5.00 |
| 10_12 | 150 | 5.00 |
| 10_13 | 150 | 5.00 |
| 10_14 | 150 | 5.00 |
| 10_15 | 150 | 5.00 |
| 10_16 | 150 | 5.00 |
| 10_17 | 150 | 5.00 |
| 10_18 | 150 | 5.00 |
| 10_19 | 150 | 5.00 |
| 10_20 | 150 | 5.00 |
| 10_21 | 150 | 5.00 |
| 10_22 | 150 | 5.00 |
| 10_23 | 150 | 5.00 |
| 10_24 | 150 | 5.00 |
| 10_25 | 150 | 5.00 |
| 10_26 | 150 | 5.00 |
| 10_27 | 150 | 5.00 |
| 10_28 | 150 | 5.00 |
| 10_29 | 150 | 5.00 |
| 10_30 | 150 | 5.00 |
| 10_31 | 150 | 5.00 |
| 10_32 | 150 | 5.00 |
| 10_33 | 150 | 5.00 |
| 10_34 | 150 | 5.00 |
| 10_35 | 150 | 5.00 |
| 10_36 | 150 | 5.00 |
| 10_37 | 150 | 5.00 |
| 10_38 | 150 | 5.00 |
| 10_39 | 150 | 5.00 |
| 10_40 | 150 | 5.00 |
| 10_41 | 150 | 5.00 |
| 10_42 | 150 | 5.00 |
| 10_43 | 150 | 5.00 |
| 10_44 | 150 | 5.00 |
| 10_45 | 150 | 5.00 |
| 10_46 | 150 | 5.00 |
| 10_47 | 186 | 6.20 |
| 11_1 | 150 | 5.00 |
| 11_2 | 150 | 5.00 |
| 11_3 | 150 | 5.00 |
| 11_4 | 150 | 5.00 |
| 11_5 | 150 | 5.00 |
| 11_6 | 150 | 5.00 |
| 11_7 | 150 | 5.00 |
| 11_8 | 150 | 5.00 |
| 11_9 | 150 | 5.00 |
| 11_10 | 150 | 5.00 |
| 11_11 | 150 | 5.00 |
| 11_12 | 150 | 5.00 |
| 11_13 | 150 | 5.00 |
| 11_14 | 150 | 5.00 |
| 11_15 | 150 | 5.00 |
| 11_16 | 150 | 5.00 |
| 11_17 | 150 | 5.00 |
| 11_18 | 150 | 5.00 |
| 11_19 | 150 | 5.00 |
| 11_20 | 150 | 5.00 |
| 11_21 | 150 | 5.00 |
| 11_22 | 150 | 5.00 |
| 11_23 | 150 | 5.00 |
| 11_24 | 150 | 5.00 |
| 11_25 | 150 | 5.00 |
| 11_26 | 150 | 5.00 |
| 11_27 | 150 | 5.00 |
| 11_28 | 150 | 5.00 |
| 11_29 | 150 | 5.00 |
| 11_30 | 150 | 5.00 |
| 11_31 | 150 | 5.00 |
| 11_32 | 182 | 6.07 |
| 12_1 | 150 | 5.00 |
| 12_2 | 150 | 5.00 |
| 12_3 | 150 | 5.00 |
| 12_4 | 150 | 5.00 |
| 12_5 | 150 | 5.00 |
| 12_6 | 150 | 5.00 |
| 12_7 | 150 | 5.00 |
| 12_8 | 150 | 5.00 |
| 12_9 | 150 | 5.00 |
| 12_10 | 150 | 5.00 |
| 12_11 | 150 | 5.00 |
| 12_12 | 150 | 5.00 |
| 12_13 | 150 | 5.00 |
| 12_14 | 150 | 5.00 |
| 12_15 | 150 | 5.00 |
| 12_16 | 150 | 5.00 |
| 12_17 | 150 | 5.00 |
| 12_18 | 150 | 5.00 |
| 12_19 | 150 | 5.00 |
| 12_20 | 150 | 5.00 |
| 12_21 | 150 | 5.00 |
| 12_22 | 150 | 5.00 |
| 12_23 | 150 | 5.00 |
| 12_24 | 150 | 5.00 |
| 12_25 | 150 | 5.00 |
| 12_26 | 150 | 5.00 |
| 12_27 | 150 | 5.00 |
| 12_28 | 150 | 5.00 |
| 12_29 | 150 | 5.00 |
| 12_30 | 150 | 5.00 |
| 12_31 | 150 | 5.00 |
| 12_32 | 150 | 5.00 |
| 12_33 | 150 | 5.00 |
| 12_34 | 150 | 5.00 |
| 12_35 | 150 | 5.00 |
| 12_36 | 150 | 5.00 |
| 12_37 | 150 | 5.00 |
| 12_38 | 150 | 5.00 |
| 12_39 | 150 | 5.00 |
| 12_40 | 150 | 5.00 |
| 12_41 | 150 | 5.00 |
| 12_42 | 150 | 5.00 |
| 12_43 | 150 | 5.00 |
| 12_44 | 150 | 5.00 |
| 12_45 | 199 | 6.63 |
| 13_1 | 150 | 5.00 |
| 13_2 | 150 | 5.00 |
| 13_3 | 150 | 5.00 |
| 13_4 | 150 | 5.00 |
| 13_5 | 150 | 5.00 |
| 13_6 | 150 | 5.00 |
| 13_7 | 150 | 5.00 |
| 13_8 | 150 | 5.00 |
| 13_9 | 150 | 5.00 |
| 13_10 | 150 | 5.00 |
| 13_11 | 150 | 5.00 |
| 13_12 | 150 | 5.00 |
| 13_13 | 150 | 5.00 |
| 13_14 | 150 | 5.00 |
| 13_15 | 150 | 5.00 |
| 13_16 | 150 | 5.00 |
| 13_17 | 150 | 5.00 |
| 13_18 | 150 | 5.00 |
| 13_19 | 150 | 5.00 |
| 13_20 | 150 | 5.00 |
| 13_21 | 150 | 5.00 |
| 13_22 | 150 | 5.00 |
| 13_23 | 150 | 5.00 |
| 13_24 | 150 | 5.00 |
| 13_25 | 150 | 5.00 |
| 13_26 | 150 | 5.00 |
| 13_27 | 150 | 5.00 |
| 13_28 | 150 | 5.00 |
| 13_29 | 150 | 5.00 |
| 13_30 | 150 | 5.00 |
| 13_31 | 150 | 5.00 |
| 13_32 | 150 | 5.00 |
| 13_33 | 150 | 5.00 |
| 13_34 | 150 | 5.00 |
| 13_35 | 150 | 5.00 |
| 13_36 | 150 | 5.00 |
| 13_37 | 150 | 5.00 |
| 13_38 | 150 | 5.00 |
| 13_39 | 150 | 5.00 |
| 13_40 | 150 | 5.00 |
| 13_41 | 150 | 5.00 |
| 13_42 | 178 | 5.93 |
================================================
FILE: docs/INFO_POSITIVE_CASES.md
================================================
# Information of Positive Cases in our SUN-SEG
## Abbreviation Illustration
- Shape
- Ip - Pedunculated
- Isp - Semipedunculated
- Is - Sessile
- IIa - Slightly elevated
- Location
- C - Cecum
- A - Ascending colon
- T - Transverse colon
- D - Descending colon
- S - Sigmoid colon
- R - Rectum
- Pathological diagnosis
- LA - Low-grade adenoma
- HA - High-grade adenoma
- HP - Hyperplastic polyp
- TSA - Traditional serrated adenoma
- SSL - Sessile serrated lesion
- IC - Invasive cancer (T1b)
## Statistic Table
| ID | Frames | Duration (s) | Shape | Size | Location | Pathological diagnosis |
|-------|-----|-------|-------------|-------|---|--------------------------------|
| 1_1 | 150 | 5.00 | Is | 6mm | C | Low-grade adenoma |
| 1_2 | 216 | 7.20 | Is | 6mm | C | Low-grade adenoma |
| 1_3 | 161 | 5.37 | Is | 6mm | C | Low-grade adenoma |
| 2_1 | 150 | 5.00 | Is | 18mm | R | High-grade adenoma |
| 2_2 | 150 | 5.00 | Is | 18mm | R | High-grade adenoma |
| 2_3 | 150 | 5.00 | Is | 18mm | R | High-grade adenoma |
| 2_4 | 150 | 5.00 | Is | 18mm | R | High-grade adenoma |
| 2_5 | 171 | 5.70 | Is | 18mm | R | High-grade adenoma |
| 2_6 | 99 | 3.30 | Is | 18mm | R | High-grade adenoma |
| 2_7 | 150 | 5.00 | Is | 18mm | R | High-grade adenoma |
| 2_8 | 181 | 6.03 | Is | 18mm | R | High-grade adenoma |
| 2_9 | 112 | 3.73 | Is | 18mm | R | High-grade adenoma |
| 3_1 | 198 | 6.60 | IIa | 3mm | A | Low-grade adenoma |
| 3_2 | 94 | 3.13 | IIa | 3mm | A | Low-grade adenoma |
| 4 | 80 | 2.67 | Is | 4mm | S | Low-grade adenoma |
| 5_1 | 187 | 6.23 | IIa | 3mm | T | Low-grade adenoma |
| 5_2 | 150 | 5.00 | IIa | 3mm | T | Low-grade adenoma |
| 5_3 | 150 | 5.00 | IIa | 3mm | T | Low-grade adenoma |
| 5_4 | 150 | 5.00 | IIa | 3mm | T | Low-grade adenoma |
| 5_5 | 118 | 3.93 | IIa | 3mm | T | Low-grade adenoma |
| 5_6 | 175 | 5.83 | IIa | 3mm | T | Low-grade adenoma |
| 6_1 | 131 | 4.37 | IIa | 3mm | S | Low-grade adenoma |
| 6_2 | 150 | 5.00 | IIa | 3mm | S | Low-grade adenoma |
| 6_3 | 210 | 7.00 | IIa | 3mm | S | Low-grade adenoma |
| 7_1 | 184 | 6.13 | IIa | 6mm | D | Low-grade adenoma |
| 7_2 | 131 | 4.37 | IIa | 6mm | D | Low-grade adenoma |
| 8_1 | 115 | 3.83 | Isp | 12mm | S | Low-grade adenoma |
| 8_2 | 141 | 4.70 | Isp | 12mm | S | Low-grade adenoma |
| 9 | 136 | 4.53 | Is | 4mm | S | Low-grade adenoma |
| 10_1 | 213 | 7.10 | IIa | 3mm | T | Low-grade adenoma |
| 10_2 | 223 | 7.43 | IIa | 3mm | T | Low-grade adenoma |
| 11 | 113 | 3.77 | IIa | 5mm | D | Low-grade adenoma |
| 12_1 | 184 | 6.13 | Is | 5mm | R | Low-grade adenoma |
| 12_2 | 248 | 8.27 | Is | 5mm | R | Low-grade adenoma |
| 12_3 | 106 | 3.53 | Is | 5mm | R | Low-grade adenoma |
| 13_1 | 98 | 3.27 | Is | 5mm | T | Low-grade adenoma |
| 13_2 | 205 | 6.83 | Is | 5mm | T | Low-grade adenoma |
| 13_3 | 176 | 5.87 | Is | 5mm | T | Low-grade adenoma |
| 14_1 | 150 | 5.00 | IIa | 3mm | S | Low-grade adenoma |
| 14_2 | 150 | 5.00 | IIa | 3mm | S | Low-grade adenoma |
| 14_3 | 378 | 12.60 | IIa | 3mm | S | Low-grade adenoma |
| 14_4 | 145 | 4.83 | IIa | 3mm | S | Low-grade adenoma |
| 14_5 | 160 | 5.33 | IIa | 3mm | S | Low-grade adenoma |
| 14_6 | 200 | 6.67 | IIa | 3mm | S | Low-grade adenoma |
| 15_1 | 235 | 7.83 | Is | 5mm | T | Low-grade adenoma |
| 15_2 | 150 | 5.00 | Is | 5mm | T | Low-grade adenoma |
| 15_3 | 102 | 3.40 | Is | 5mm | T | Low-grade adenoma |
| 16 | 199 | 6.63 | Is | 4mm | T | Low-grade adenoma |
| 17_1 | 126 | 4.20 | Is | 4mm | S | Low-grade adenoma |
| 17_2 | 178 | 5.93 | Is | 4mm | S | Low-grade adenoma |
| 18_1 | 91 | 3.03 | Is | 2mm | S | Hyperplastic polyp |
| 18_2 | 152 | 5.07 | Is | 2mm | S | Hyperplastic polyp |
| 19 | 96 | 3.20 | IIa | 3mm | T | Low-grade adenoma |
| 20_1 | 150 | 5.00 | IIa | 3mm | A | Low-grade adenoma |
| 20_2 | 205 | 6.83 | IIa | 3mm | A | Low-grade adenoma |
| 20_3 | 148 | 4.93 | IIa | 3mm | A | Low-grade adenoma |
| 20_4 | 150 | 5.00 | IIa | 3mm | A | Low-grade adenoma |
| 20_5 | 191 | 6.37 | IIa | 3mm | A | Low-grade adenoma |
| 20_6 | 150 | 5.00 | IIa | 3mm | A | Low-grade adenoma |
| 20_7 | 170 | 5.67 | IIa | 3mm | A | Low-grade adenoma |
| 20_8 | 233 | 7.77 | IIa | 3mm | A | Low-grade adenoma |
| 20_9 | 150 | 5.00 | IIa | 3mm | A | Low-grade adenoma |
| 20_10 | 150 | 5.00 | IIa | 3mm | A | Low-grade adenoma |
| 20_11 | 150 | 5.00 | IIa | 3mm | A | Low-grade adenoma |
| 20_12 | 156 | 5.20 | IIa | 3mm | A | Low-grade adenoma |
| 20_13 | 150 | 5.00 | IIa | 3mm | A | Low-grade adenoma |
| 20_14 | 150 | 5.00 | IIa | 3mm | A | Low-grade adenoma |
| 20_15 | 150 | 5.00 | IIa | 3mm | A | Low-grade adenoma |
| 20_16 | 150 | 5.00 | IIa | 3mm | A | Low-grade adenoma |
| 20_17 | 150 | 5.00 | IIa | 3mm | A | Low-grade adenoma |
| 20_18 | 150 | 5.00 | IIa | 3mm | A | Low-grade adenoma |
| 20_19 | 118 | 3.93 | IIa | 3mm | A | Low-grade adenoma |
| 20_20 | 138 | 4.60 | IIa | 3mm | A | Low-grade adenoma |
| 21 | 100 | 3.33 | IIa | 3mm | S | Low-grade adenoma |
| 22 | 314 | 10.47 | IIa | 2mm | A | Low-grade adenoma |
| 23 | 182 | 6.07 | Ip | 12mm | A | Low-grade adenoma |
| 24_1 | 150 | 5.00 | Ip | 15mm- | S | Low-grade adenoma |
| 24_2 | 150 | 5.00 | Ip | 15mm- | S | Low-grade adenoma |
| 24_3 | 152 | 5.07 | Ip | 15mm- | S | Low-grade adenoma |
| 24_4 | 150 | 5.00 | Ip | 15mm- | S | Low-grade adenoma |
| 24_5 | 150 | 5.00 | Ip | 15mm- | S | Low-grade adenoma |
| 24_6 | 221 | 7.37 | Ip | 15mm- | S | Low-grade adenoma |
| 25_1 | 202 | 6.73 | Is | 7mm | S | Low-grade adenoma |
| 25_2 | 136 | 4.53 | Is | 7mm | S | Low-grade adenoma |
| 26_1 | 151 | 5.03 | Is | 5mm | D | Low-grade adenoma |
| 26_2 | 219 | 7.30 | Is | 5mm | D | Low-grade adenoma |
| 27 | 249 | 8.30 | Is | 5mm | A | Hyperplastic polyp |
| 28 | 195 | 6.50 | Is | 2mm | T | Low-grade adenoma |
| 29_1 | 214 | 7.13 | Isp | 13mm | S | Low-grade adenoma |
| 29_2 | 163 | 5.43 | Isp | 13mm | S | Low-grade adenoma |
| 30 | 224 | 7.47 | IIa | 4mm | S | Low-grade adenoma |
| 31 | 183 | 6.10 | Ip | 12mm | D | Low-grade adenoma |
| 32_1 | 210 | 7.00 | Ip | 15mm- | A | Traditional serrated adenoma |
| 32_2 | 100 | 3.33 | Ip | 15mm- | A | Traditional serrated adenoma |
| 32_3 | 165 | 5.50 | Ip | 15mm- | A | Traditional serrated adenoma |
| 32_4 | 185 | 6.17 | Ip | 15mm- | A | Traditional serrated adenoma |
| 32_5 | 150 | 5.00 | Ip | 15mm- | A | Traditional serrated adenoma |
| 32_6 | 171 | 5.70 | Ip | 15mm- | A | Traditional serrated adenoma |
| 33_1 | 207 | 6.90 | Is | 5mm | S | Low-grade adenoma |
| 33_2 | 186 | 6.20 | Is | 5mm | S | Low-grade adenoma |
| 33_3 | 201 | 6.70 | Is | 5mm | S | Low-grade adenoma |
| 34_1 | 121 | 4.03 | Is | 3mm | A | Low-grade adenoma |
| 34_2 | 124 | 4.13 | Is | 3mm | A | Low-grade adenoma |
| 35_1 | 185 | 6.17 | Ip | 15mm- | S | High-grade adenoma |
| 35_2 | 90 | 3.00 | Ip | 15mm- | S | High-grade adenoma |
| 35_3 | 193 | 6.43 | Ip | 15mm- | S | High-grade adenoma |
| 35_4 | 223 | 7.43 | Ip | 15mm- | S | High-grade adenoma |
| 35_5 | 150 | 5.00 | Ip | 15mm- | S | High-grade adenoma |
| 35_6 | 131 | 4.37 | Ip | 15mm- | S | High-grade adenoma |
| 35_7 | 124 | 4.13 | Ip | 15mm- | S | High-grade adenoma |
| 35_8 | 116 | 3.87 | Ip | 15mm- | S | High-grade adenoma |
| 36_1 | 334 | 11.13 | IIa | 7mm | S | Low-grade adenoma |
| 36_2 | 150 | 5.00 | IIa | 7mm | S | Low-grade adenoma |
| 36_3 | 203 | 6.77 | IIa | 7mm | S | Low-grade adenoma |
| 36_4 | 128 | 4.27 | IIa | 7mm | S | Low-grade adenoma |
| 37_1 | 187 | 6.23 | Is | 7mm | T | Low-grade adenoma |
| 37_2 | 159 | 5.30 | Is | 7mm | T | Low-grade adenoma |
| 37_3 | 102 | 3.40 | Is | 7mm | T | Low-grade adenoma |
| 38_1 | 255 | 8.50 | Is | 5mm | A | Low-grade adenoma |
| 38_2 | 254 | 8.47 | Is | 5mm | A | Low-grade adenoma |
| 39_1 | 294 | 9.80 | IIa | 13mm | A | Low-grade adenoma |
| 39_2 | 270 | 9.00 | IIa | 13mm | A | Low-grade adenoma |
| 39_3 | 149 | 4.97 | IIa | 13mm | A | Low-grade adenoma |
| 40 | 159 | 5.30 | IIa | 5mm | T | Low-grade adenoma |
| 41 | 108 | 3.60 | IIa | 3mm | R | Low-grade adenoma |
| 42 | 268 | 8.93 | Is | 7mm | T | Low-grade adenoma |
| 43 | 260 | 8.67 | Isp | 10mm | A | Low-grade adenoma |
| 44_1 | 150 | 5.00 | IIa | 5mm | S | Low-grade adenoma |
| 44_2 | 150 | 5.00 | IIa | 5mm | S | Low-grade adenoma |
| 44_3 | 217 | 7.23 | IIa | 5mm | S | Low-grade adenoma |
| 44_4 | 228 | 7.60 | IIa | 5mm | S | Low-grade adenoma |
| 45_1 | 118 | 3.93 | Is | 3mm | A | Low-grade adenoma |
| 45_2 | 134 | 4.47 | Is | 3mm | A | Low-grade adenoma |
| 45_3 | 131 | 4.37 | Is | 3mm | A | Low-grade adenoma |
| 46 | 170 | 5.67 | IIa | 2mm | T | Hyperplastic polyp |
| 47_1 | 150 | 5.00 | Is | 5mm | T | Low-grade adenoma |
| 47_2 | 150 | 5.00 | Is | 5mm | T | Low-grade adenoma |
| 47_3 | 150 | 5.00 | Is | 5mm | T | Low-grade adenoma |
| 47_4 | 133 | 4.43 | Is | 5mm | T | Low-grade adenoma |
| 47_5 | 122 | 4.07 | Is | 5mm | T | Low-grade adenoma |
| 48 | 176 | 5.87 | Is | 3mm | T | Low-grade adenoma |
| 49 | 181 | 6.03 | IIa | 3mm | T | Low-grade adenoma |
| 50_1 | 203 | 6.77 | Ip | 10mm | S | Low-grade adenoma |
| 50_2 | 150 | 5.00 | Ip | 10mm | S | Low-grade adenoma |
| 50_3 | 150 | 5.00 | Ip | 10mm | S | Low-grade adenoma |
| 50_4 | 237 | 7.90 | Ip | 10mm | S | Low-grade adenoma |
| 51_1 | 175 | 5.83 | IIa(LST-NG) | 15mm- | C | Low-grade adenoma |
| 51_2 | 250 | 8.33 | IIa(LST-NG) | 15mm- | C | Low-grade adenoma |
| 51_3 | 210 | 7.00 | IIa(LST-NG) | 15mm- | C | Low-grade adenoma |
| 51_4 | 150 | 5.00 | IIa(LST-NG) | 15mm- | C | Low-grade adenoma |
| 51_5 | 150 | 5.00 | IIa(LST-NG) | 15mm- | C | Low-grade adenoma |
| 51_6 | 150 | 5.00 | IIa(LST-NG) | 15mm- | C | Low-grade adenoma |
| 51_7 | 134 | 4.47 | IIa(LST-NG) | 15mm- | C | Low-grade adenoma |
| 51_8 | 150 | 5.00 | IIa(LST-NG) | 15mm- | C | Low-grade adenoma |
| 51_9 | 150 | 5.00 | IIa(LST-NG) | 15mm- | C | Low-grade adenoma |
| 51_10 | 218 | 7.27 | IIa(LST-NG) | 15mm- | C | Low-grade adenoma |
| 52 | 207 | 6.90 | IIa | 6mm | S | Low-grade adenoma |
| 53 | 245 | 8.17 | Is | 4mm | R | Hyperplastic polyp |
| 54_1 | 211 | 7.03 | Is | 4mm | S | Low-grade adenoma |
| 54_2 | 134 | 4.47 | Is | 4mm | S | Low-grade adenoma |
| 55_1 | 237 | 7.90 | Is | 3mm | A | Low-grade adenoma |
| 55_2 | 336 | 11.20 | Is | 3mm | A | Low-grade adenoma |
| 55_3 | 127 | 4.23 | Is | 3mm | A | Low-grade adenoma |
| 56_1 | 150 | 5.00 | Is | 4mm | S | Hyperplastic polyp |
| 56_2 | 98 | 3.27 | Is | 4mm | S | Hyperplastic polyp |
| 57 | 326 | 10.87 | Is | 5mm | T | Low-grade adenoma |
| 58 | 267 | 8.90 | IIa | 6mm | T | Sessile serrated lesion |
| 59_1 | 281 | 9.37 | Isp | 8mm | S | Traditional serrated adenoma |
| 59_2 | 150 | 5.00 | Isp | 8mm | S | Traditional serrated adenoma |
| 59_3 | 215 | 7.17 | Isp | 8mm | S | Traditional serrated adenoma |
| 60 | 146 | 4.87 | IIa | 8mm | T | Low-grade adenoma |
| 61_1 | 259 | 8.63 | Isp | 6mm | A | Low-grade adenoma |
| 61_2 | 199 | 6.63 | Isp | 6mm | A | Low-grade adenoma |
| 61_3 | 221 | 7.37 | Isp | 6mm | A | Low-grade adenoma |
| 62_1 | 190 | 6.33 | Is | 7mm | A | Low-grade adenoma |
| 62_2 | 161 | 5.37 | Is | 7mm | A | Low-grade adenoma |
| 63_1 | 326 | 10.87 | Is | 7mm | R | Invasive cancer (T1b) |
| 63_2 | 133 | 4.43 | Is | 7mm | R | Invasive cancer (T1b) |
| 63_3 | 173 | 5.77 | Is | 7mm | R | Invasive cancer (T1b) |
| 64 | 81 | 2.70 | IIa | 3mm | S | Low-grade adenoma |
| 65 | 222 | 7.40 | IIa | 3mm | C | Low-grade adenoma |
| 66_1 | 165 | 5.50 | Is | 6mm | S | Low-grade adenoma |
| 66_2 | 323 | 10.77 | Is | 6mm | S | Low-grade adenoma |
| 66_3 | 150 | 5.00 | Is | 6mm | S | Low-grade adenoma |
| 66_4 | 236 | 7.87 | Is | 6mm | S | Low-grade adenoma |
| 66_5 | 150 | 5.00 | Is | 6mm | S | Low-grade adenoma |
| 66_6 | 134 | 4.47 | Is | 6mm | S | Low-grade adenoma |
| 66_7 | 177 | 5.90 | Is | 6mm | S | Low-grade adenoma |
| 66_8 | 150 | 5.00 | Is | 6mm | S | Low-grade adenoma |
| 66_9 | 200 | 6.67 | Is | 6mm | S | Low-grade adenoma |
| 67 | 191 | 6.37 | IIa | 5mm | T | Low-grade adenoma |
| 68_1 | 164 | 5.47 | Is | 15mm- | R | High-grade adenoma |
| 68_2 | 180 | 6.00 | Is | 15mm- | R | High-grade adenoma |
| 68_3 | 163 | 5.43 | Is | 15mm- | R | High-grade adenoma |
| 68_4 | 228 | 7.60 | Is | 15mm- | R | High-grade adenoma |
| 68_5 | 150 | 5.00 | Is | 15mm- | R | High-grade adenoma |
| 68_6 | 204 | 6.80 | Is | 15mm- | R | High-grade adenoma |
| 68_7 | 230 | 7.67 | Is | 15mm- | R | High-grade adenoma |
| 69 | 130 | 4.33 | IIa | 3mm | D | Low-grade adenoma |
| 70 | 264 | 8.80 | Ip | 15mm- | S | Low-grade adenoma |
| 71_1 | 128 | 4.27 | Is | 4mm | A | Low-grade adenoma |
| 71_2 | 150 | 5.00 | Is | 4mm | A | Low-grade adenoma |
| 71_3 | 150 | 5.00 | Is | 4mm | A | Low-grade adenoma |
| 71_4 | 144 | 4.80 | Is | 4mm | A | Low-grade adenoma |
| 71_5 | 226 | 7.53 | Is | 4mm | A | Low-grade adenoma |
| 71_6 | 223 | 7.43 | Is | 4mm | A | Low-grade adenoma |
| 72_1 | 221 | 7.37 | Is | 5mm | A | Low-grade adenoma |
| 72_2 | 234 | 7.80 | Is | 5mm | A | Low-grade adenoma |
| 72_3 | 104 | 3.47 | Is | 5mm | A | Low-grade adenoma |
| 72_4 | 215 | 7.17 | Is | 5mm | A | Low-grade adenoma |
| 73_1 | 161 | 5.37 | Is | 3mm | C | Low-grade adenoma |
| 73_2 | 111 | 3.70 | Is | 3mm | C | Low-grade adenoma |
| 73_3 | 164 | 5.47 | Is | 3mm | C | Low-grade adenoma |
| 73_4 | 128 | 4.27 | Is | 3mm | C | Low-grade adenoma |
| 73_5 | 150 | 5.00 | Is | 3mm | C | Low-grade adenoma |
| 73_6 | 169 | 5.63 | Is | 3mm | C | Low-grade adenoma |
| 73_7 | 150 | 5.00 | Is | 3mm | C | Low-grade adenoma |
| 73_8 | 150 | 5.00 | Is | 3mm | C | Low-grade adenoma |
| 73_9 | 102 | 3.40 | Is | 3mm | C | Low-grade adenoma |
| 74_1 | 157 | 5.23 | Isp | 5mm | S | Low-grade adenoma |
| 74_2 | 119 | 3.97 | Isp | 5mm | S | Low-grade adenoma |
| 75_1 | 150 | 5.00 | Is | 3mm | T | Low-grade adenoma |
| 75_2 | 193 | 6.43 | Is | 3mm | T | Low-grade adenoma |
| 76_1 | 162 | 5.40 | Is | 3mm | C | Low-grade adenoma |
| 76_2 | 181 | 6.03 | Is | 3mm | C | Low-grade adenoma |
| 77 | 215 | 7.17 | Is | 4mm | A | Low-grade adenoma |
| 78_1 | 170 | 5.67 | Isp | 12mm | S | High-grade adenoma |
| 78_2 | 97 | 3.23 | Isp | 12mm | S | High-grade adenoma |
| 79 | 76 | 2.53 | Is | 4mm | D | Low-grade adenoma |
| 80_1 | 160 | 5.33 | Is | 10mm | S | Low-grade adenoma |
| 80_2 | 150 | 5.00 | Is | 10mm | S | Low-grade adenoma |
| 80_3 | 150 | 5.00 | Is | 10mm | S | Low-grade adenoma |
| 80_4 | 150 | 5.00 | Is | 10mm | S | Low-grade adenoma |
| 80_5 | 137 | 4.57 | Is | 10mm | S | Low-grade adenoma |
| 80_6 | 150 | 5.00 | Is | 10mm | S | Low-grade adenoma |
| 80_7 | 150 | 5.00 | Is | 10mm | S | Low-grade adenoma |
| 80_8 | 145 | 4.83 | Is | 10mm | S | Low-grade adenoma |
| 81_1 | 153 | 5.10 | Is | 6mm | S | Low-grade adenoma |
| 81_2 | 274 | 9.13 | Is | 6mm | S | Low-grade adenoma |
| 82 | 111 | 3.70 | IIa | 3mm | S | Sessile serrated lesion |
| 83_1 | 283 | 9.43 | Isp | 13mm | R | Low-grade adenoma |
| 83_2 | 378 | 12.60 | Isp | 13mm | R | Low-grade adenoma |
| 83_3 | 134 | 4.47 | Isp | 13mm | R | Low-grade adenoma |
| 84 | 218 | 7.27 | Is | 5mm | D | Low-grade adenoma |
| 85_1 | 150 | 5.00 | IIa | 8mm | A | Low-grade adenoma |
| 85_2 | 90 | 3.00 | IIa | 8mm | A | Low-grade adenoma |
| 85_3 | 197 | 6.57 | IIa | 8mm | A | Low-grade adenoma |
| 85_4 | 197 | 6.57 | IIa | 8mm | A | Low-grade adenoma |
| 85_5 | 239 | 7.97 | IIa | 8mm | A | Low-grade adenoma |
| 85_6 | 150 | 5.00 | IIa | 8mm | A | Low-grade adenoma |
| 85_7 | 150 | 5.00 | IIa | 8mm | A | Low-grade adenoma |
| 85_8 | 220 | 7.33 | IIa | 8mm | A | Low-grade adenoma |
| 86_1 | 150 | 5.00 | IIa | 4mm | S | Low-grade adenoma |
| 86_2 | 107 | 3.57 | IIa | 4mm | S | Low-grade adenoma |
| 87_1 | 250 | 8.33 | Is | 3mm | C | Low-grade adenoma |
| 87_2 | 204 | 6.80 | Is | 3mm | C | Low-grade adenoma |
| 88_1 | 249 | 8.30 | Is | 4mm | A | Low-grade adenoma |
| 89 | 149 | 4.97 | Ip | 5mm | D | Low-grade adenoma |
| 90_1 | 150 | 5.00 | Is | 10mm | A | Sessile serrated lesion |
| 90_2 | 150 | 5.00 | Is | 10mm | A | Sessile serrated lesion |
| 90_3 | 179 | 5.97 | Is | 10mm | A | Sessile serrated lesion |
| 91_1 | 150 | 5.00 | IIa | 13mm | A | Low-grade adenoma |
| 91_2 | 150 | 5.00 | IIa | 13mm | A | Low-grade adenoma |
| 91_3 | 150 | 5.00 | IIa | 13mm | A | Low-grade adenoma |
| 91_4 | 109 | 3.63 | IIa | 13mm | A | Low-grade adenoma |
| 91_5 | 150 | 5.00 | IIa | 13mm | A | Low-grade adenoma |
| 91_6 | 118 | 3.93 | IIa | 13mm | A | Low-grade adenoma |
| 91_7 | 234 | 7.80 | IIa | 13mm | A | Low-grade adenoma |
| 92_1 | 201 | 6.70 | Is | 7mm | D | Low-grade adenoma |
| 92_2 | 190 | 6.33 | Is | 7mm | D | Low-grade adenoma |
| 93_1 | 217 | 7.23 | Is | 7mm | D | Low-grade adenoma |
| 93_2 | 235 | 7.83 | Is | 7mm | D | Low-grade adenoma |
| 94 | 136 | 4.53 | Is | 6mm | S | Low-grade adenoma |
| 95_1 | 259 | 8.63 | Isp | 8mm | S | Low-grade adenoma |
| 95_2 | 102 | 3.40 | Isp | 8mm | S | Low-grade adenoma |
| 95_3 | 150 | 5.00 | Isp | 8mm | S | Low-grade adenoma |
| 95_4 | 95 | 3.17 | Isp | 8mm | S | Low-grade adenoma |
| 96_1 | 115 | 3.83 | Is | 5mm | S | Hyperplastic polyp |
| 96_2 | 186 | 6.20 | Is | 5mm | S | Hyperplastic polyp |
| 97_1 | 150 | 5.00 | IIa | 15mm- | C | Sessile serrated lesion |
| 97_2 | 154 | 5.13 | IIa | 15mm- | C | Sessile serrated lesion |
| 97_3 | 127 | 4.23 | IIa | 15mm- | C | Sessile serrated lesion |
| 98 | 170 | 5.67 | IIa | 4mm | T | Low-grade adenoma |
| 99 | 161 | 5.37 | Is | 5mm | S | Low-grade adenoma |
| 100 | 188 | 6.27 | IIa | 3mm | R | Hyperplastic polyp |
================================================
FILE: docs/RELEASE_NOTES.md
================================================
- We greatly appreciate [@Yuli Zhou](https://github.com/zhoustan) for the feedback and for fixing the sorting bug in our test and evaluation processes.
This bug fix has a very minor impact on performance. Below, we showcase the results before and after the fix:
| Dataset | Method | Status | Smeasure | meanEm | wFmeasure | maxDice |
|------------------------|------------------|----------------|----------|--------|-----------|---------|
| TestEasyDataset-Seen | 2022-MIR-PNSPlus | before-fix | 0.917 | 0.924 | 0.848 | 0.888 |
| TestEasyDataset-Seen | 2022-MIR-PNSPlus | after-fix | 0.917 | 0.924 | 0.848 | 0.888 |
| TestHardDataset-Seen | 2022-MIR-PNSPlus | before-fix | 0.887 | 0.929 | 0.806 | 0.855 |
| TestHardDataset-Seen | 2022-MIR-PNSPlus | after-fix | 0.887 | 0.902 | 0.806 | 0.855 |
| Dataset | Method | Status | Smeasure | meanEm | wFmeasure | maxDice | meanFm | meanSen |
|------------------------|------------------|----------------|----------|--------|-----------|---------|--------|---------|
| TestEasyDataset-Unseen | 2022-MIR-PNSPlus | before-fix | 0.806 | 0.798 | 0.676 | 0.756 | 0.730 | 0.630 |
| TestEasyDataset-Unseen | 2022-MIR-PNSPlus | after-fix | 0.806 | 0.798 | 0.676 | 0.756 | 0.730 | 0.630 |
| TestHardDataset-Unseen | 2022-MIR-PNSPlus | before-fix | 0.797 | 0.793 | 0.653 | 0.737 | 0.709 | 0.623 |
| TestHardDataset-Unseen | 2022-MIR-PNSPlus | after-fix | 0.798 | 0.793 | 0.654 | 0.737 | 0.709 | 0.624 |
Reminder: To ensure the correctness of results, we strongly recommend running the sorting strategy on a Linux system.
================================================
FILE: eval/README.md
================================================
# VPS Evaluation Toolbox
This toolbox is used to evaluate the performance of video polyp segmentation task.
# Usage
- Prerequisites of environment:
```bash
python -m pip install opencv-python tdqm prettytable scikit-learn
```
- Running the evaluation:
```bash
sh ./eval.sh
```
By running the script, results of all models on SUN-SEG dataset will be evaluated simultaneously. If you want to evaluate the specific models, please modify the `$MODEL_NAMES`variable in `eval.sh` which is corresponding to the argument `--model_lst`. Note that the modified model name should be the same to the folder name under `./data/Pred/`.
In `vps_evaluator.py`, you can specify `--metric_list` to decide the applying metrics. `--txt_name` denotes the folder name of evaluation result. `--data_lst` and `--check_integrity` represent the used dataset and the integrity examination of result maps and ground truth.
# Citation
If you have found our work useful, please use the following reference to cite this project:
@article{ji2022video,
title={Video polyp segmentation: A deep learning perspective},
author={Ji, Ge-Peng and Xiao, Guobao and Chou, Yu-Cheng and Fan, Deng-Ping and Zhao, Kai and Chen, Geng and Van Gool, Luc},
journal={Machine Intelligence Research},
volume={19},
number={6},
pages={531--549},
year={2022},
publisher={Springer}
}
@inproceedings{ji2021progressively,
title={Progressively normalized self-attention network for video polyp segmentation},
author={Ji, Ge-Peng and Chou, Yu-Cheng and Fan, Deng-Ping and Chen, Geng and Fu, Huazhu and Jha, Debesh and Shao, Ling},
booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
pages={142--152},
year={2021},
organization={Springer}
}
@inproceedings{fan2020pranet,
title={Pranet: Parallel reverse attention network for polyp segmentation},
author={Fan, Deng-Ping and Ji, Ge-Peng and Zhou, Tao and Chen, Geng and Fu, Huazhu and Shen, Jianbing and Shao, Ling},
booktitle={International conference on medical image computing and computer-assisted intervention},
pages={263--273},
year={2020},
organization={Springer}
}
================================================
FILE: eval/eval-result/2015-MICCAI-UNet/TestEasyDataset_eval.txt
================================================
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
| TestEasyDataset | 2015-MICCAI-UNet | 0.720 | 0.810 | 0.543 | 0.625 | 0.606 | 0.508 |
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2015-MICCAI-UNet/TestHardDataset_eval.txt
================================================
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
| TestHardDataset | 2015-MICCAI-UNet | 0.710 | 0.801 | 0.524 | 0.619 | 0.602 | 0.493 |
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2018-TMI-UNet++/TestEasyDataset_eval.txt
================================================
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
| TestEasyDataset | 2018-TMI-UNet++ | 0.731 | 0.803 | 0.560 | 0.633 | 0.620 | 0.524 |
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2018-TMI-UNet++/TestHardDataset_eval.txt
================================================
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
| TestHardDataset | 2018-TMI-UNet++ | 0.714 | 0.795 | 0.526 | 0.605 | 0.595 | 0.488 |
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2019-TPAMI-COSNet/TestEasyDataset_eval.txt
================================================
+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+
| TestEasyDataset | 2019-TPAMI-COSNet | 0.707 | 0.851 | 0.513 | 0.667 | 0.649 | 0.548 |
+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2019-TPAMI-COSNet/TestHardDataset_eval.txt
================================================
+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+
| TestHardDataset | 2019-TPAMI-COSNet | 0.706 | 0.845 | 0.501 | 0.651 | 0.639 | 0.531 |
+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2020-AAAI-PCSA/TestEasyDataset_eval.txt
================================================
+-----------------+----------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+----------------+----------+-------+-----------+-------+---------+--------+
| TestEasyDataset | 2020-AAAI-PCSA | 0.728 | 0.845 | 0.515 | 0.649 | 0.631 | 0.518 |
+-----------------+----------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2020-AAAI-PCSA/TestHardDataset_eval.txt
================================================
+-----------------+----------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+----------------+----------+-------+-----------+-------+---------+--------+
| TestHardDataset | 2020-AAAI-PCSA | 0.711 | 0.823 | 0.481 | 0.624 | 0.609 | 0.489 |
+-----------------+----------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2020-MICCAI-23DCNN/TestEasyDataset_eval.txt
================================================
+-----------------+--------------------+----------+-------+-----------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxDice | maxIoU |
+-----------------+--------------------+----------+-------+-----------+---------+--------+
| TestEasyDataset | 2020-MICCAI-23DCNN | 0.817 | 0.872 | 0.698 | 0.755 | 0.668 |
+-----------------+--------------------+----------+-------+-----------+---------+--------+
================================================
FILE: eval/eval-result/2020-MICCAI-23DCNN/TestHardDataset_eval.txt
================================================
+-----------------+--------------------+----------+-------+-----------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxDice | maxIoU |
+-----------------+--------------------+----------+-------+-----------+---------+--------+
| TestHardDataset | 2020-MICCAI-23DCNN | 0.806 | 0.863 | 0.671 | 0.737 | 0.643 |
+-----------------+--------------------+----------+-------+-----------+---------+--------+
================================================
FILE: eval/eval-result/2020-MICCAI-ACSNet/TestEasyDataset_eval.txt
================================================
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
| TestEasyDataset | 2020-MICCAI-ACSNet | 0.820 | 0.889 | 0.702 | 0.773 | 0.760 | 0.682 |
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2020-MICCAI-ACSNet/TestHardDataset_eval.txt
================================================
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
| TestHardDataset | 2020-MICCAI-ACSNet | 0.818 | 0.885 | 0.696 | 0.769 | 0.758 | 0.673 |
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2020-MICCAI-PraNet/TestEasyDataset_eval.txt
================================================
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
| TestEasyDataset | 2020-MICCAI-PraNet | 0.780 | 0.877 | 0.649 | 0.710 | 0.689 | 0.608 |
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2020-MICCAI-PraNet/TestHardDataset_eval.txt
================================================
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
| TestHardDataset | 2020-MICCAI-PraNet | 0.759 | 0.846 | 0.615 | 0.683 | 0.660 | 0.569 |
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2020-TIP-MATNet/TestEasyDataset_eval.txt
================================================
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
| TestEasyDataset | 2020-TIP-MATNet | 0.800 | 0.889 | 0.618 | 0.757 | 0.739 | 0.632 |
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2020-TIP-MATNet/TestHardDataset_eval.txt
================================================
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
| TestHardDataset | 2020-TIP-MATNet | 0.802 | 0.881 | 0.601 | 0.745 | 0.728 | 0.616 |
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2021-ICCV-DCFNet/TestEasyDataset_eval.txt
================================================
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
| TestEasyDataset | 2021-ICCV-DCFNet | 0.536 | 0.560 | 0.294 | 0.353 | 0.344 | 0.274 |
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2021-ICCV-DCFNet/TestHardDataset_eval.txt
================================================
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
| TestHardDataset | 2021-ICCV-DCFNet | 0.542 | 0.575 | 0.301 | 0.363 | 0.357 | 0.284 |
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2021-ICCV-FSNet/TestEasyDataset_eval.txt
================================================
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
| TestEasyDataset | 2021-ICCV-FSNet | 0.771 | 0.842 | 0.625 | 0.769 | 0.747 | 0.658 |
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2021-ICCV-FSNet/TestHardDataset_eval.txt
================================================
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
| TestHardDataset | 2021-ICCV-FSNet | 0.763 | 0.849 | 0.608 | 0.752 | 0.738 | 0.641 |
+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2021-MICCAI-PNSNet/TestEasyDataset_eval.txt
================================================
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
| TestEasyDataset | 2021-MICCAI-PNSNet | 0.805 | 0.843 | 0.677 | 0.738 | 0.724 | 0.649 |
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2021-MICCAI-PNSNet/TestHardDataset_eval.txt
================================================
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
| TestHardDataset | 2021-MICCAI-PNSNet | 0.799 | 0.849 | 0.665 | 0.728 | 0.719 | 0.637 |
+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2021-MICCAI-SANet/TestEasyDataset_eval.txt
================================================
+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+
| TestEasyDataset | 2021-MICCAI-SANet | 0.755 | 0.870 | 0.621 | 0.707 | 0.693 | 0.595 |
+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2021-MICCAI-SANet/TestHardDataset_eval.txt
================================================
+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+
| TestHardDataset | 2021-MICCAI-SANet | 0.736 | 0.849 | 0.577 | 0.647 | 0.640 | 0.543 |
+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2021-NIPS-AMD/TestEasyDataset_eval.txt
================================================
+-----------------+---------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+---------------+----------+-------+-----------+-------+---------+--------+
| TestEasyDataset | 2021-NIPS-AMD | 0.473 | 0.602 | 0.127 | 0.219 | 0.260 | 0.165 |
+-----------------+---------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2021-NIPS-AMD/TestHardDataset_eval.txt
================================================
+-----------------+---------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+---------------+----------+-------+-----------+-------+---------+--------+
| TestHardDataset | 2021-NIPS-AMD | 0.474 | 0.585 | 0.124 | 0.209 | 0.244 | 0.155 |
+-----------------+---------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2022-TMI-PNSPlus/TestEasyDataset_eval.txt
================================================
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
| TestEasyDataset | 2022-MIR-PNSPlus | 0.837 | 0.910 | 0.723 | 0.803 | 0.787 | 0.704 |
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval-result/2022-TMI-PNSPlus/TestHardDataset_eval.txt
================================================
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
| Dataset | Method | Smeasure | maxEm | wFmeasure | maxFm | maxDice | maxIoU |
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
| TestHardDataset | 2022-MIR-PNSPlus | 0.826 | 0.891 | 0.701 | 0.785 | 0.770 | 0.679 |
+-----------------+------------------+----------+-------+-----------+-------+---------+--------+
================================================
FILE: eval/eval.sh
================================================
# The candidate competitors are listed here:
# MODEL_NAMES=('2022-MIR-PNSPlus' '2021-NIPS-AMD' '2021-MICCAI-PNSNet' '2021-ICCV-FSNet' '2021-ICCV-DCFNet' '2020-TIP-MATNet' '2020-MICCAI-23DCNN' '2020-AAAI-PCSA' '2019-TPAMI-COSNet' '2021-TPAMI-SINetV2' '2021-MICCAI-SANet' '2020-MICCAI-PraNet' '2020-MICCAI-ACSNet' '2018-TMI-UNet++' '2015-MICCAI-UNet')
MODEL_NAMES=('2022-MIR-PNSPlus')
for MODEL_NAME in ${MODEL_NAMES[*]}
do
nohup python -u vps_evaluator.py --data_lst TestEasyDataset/Unseen --model_lst $MODEL_NAME --txt_name $MODEL_NAME >> ./loggings/$MODEL_NAME-1.log &
nohup python -u vps_evaluator.py --data_lst TestEasyDataset/Seen --model_lst $MODEL_NAME --txt_name $MODEL_NAME >> ./loggings/$MODEL_NAME-2.log &
nohup python -u vps_evaluator.py --data_lst TestHardDataset/Unseen --model_lst $MODEL_NAME --txt_name $MODEL_NAME >> ./loggings/$MODEL_NAME-2.log &
nohup python -u vps_evaluator.py --data_lst TestHardDataset/Seen --model_lst $MODEL_NAME --txt_name $MODEL_NAME >> ./loggings/$MODEL_NAME-2.log &
done
================================================
FILE: eval/metrics.py
================================================
# -*- coding: utf-8 -*-
# @Time : 2021/09/13
# @Author : Johnson-Chou
# @Email : johnson111788@gmail.com
# @FileName : metrics.py
# @Reference: https://github.com/mczhuge/SOCToolbox
import numpy as np
from scipy.ndimage import convolve, distance_transform_edt as bwdist
_EPS = np.spacing(1)
_TYPE = np.float64
def _prepare_data(pred: np.ndarray, gt: np.ndarray) -> tuple:
gt = gt > 128
pred = pred / 255
if pred.max() != pred.min():
pred = (pred - pred.min()) / (pred.max() - pred.min())
return pred, gt
def _get_adaptive_threshold(matrix: np.ndarray, max_value: float = 1) -> float:
return min(2 * matrix.mean(), max_value)
class Fmeasure(object):
def __init__(self, length, beta: float = 0.3):
self.beta = beta
self.precisions = []
self.recalls = []
self.adaptive_fms = []
self.changeable_fms = []
def step(self, pred: np.ndarray, gt: np.ndarray, idx):
pred, gt = _prepare_data(pred, gt)
adaptive_fm = self.cal_adaptive_fm(pred=pred, gt=gt)
self.adaptive_fms.append(adaptive_fm)
precisions, recalls, changeable_fms = self.cal_pr(pred=pred, gt=gt)
self.precisions.append(precisions)
self.recalls.append(recalls)
self.changeable_fms.append(changeable_fms)
def cal_adaptive_fm(self, pred: np.ndarray, gt: np.ndarray) -> float:
adaptive_threshold = _get_adaptive_threshold(pred, max_value=1)
binary_predcition = pred >= adaptive_threshold
area_intersection = binary_predcition[gt].sum()
if area_intersection == 0:
adaptive_fm = 0
else:
pre = area_intersection / np.count_nonzero(binary_predcition)
rec = area_intersection / np.count_nonzero(gt)
# F_beta measure
adaptive_fm = (1 + self.beta) * pre * rec / (self.beta * pre + rec)
return adaptive_fm
def cal_pr(self, pred: np.ndarray, gt: np.ndarray) -> tuple:
pred = (pred * 255).astype(np.uint8)
bins = np.linspace(0, 256, 257)
fg_hist, _ = np.histogram(pred[gt], bins=bins)
bg_hist, _ = np.histogram(pred[~gt], bins=bins)
fg_w_thrs = np.cumsum(np.flip(fg_hist), axis=0)
bg_w_thrs = np.cumsum(np.flip(bg_hist), axis=0)
TPs = fg_w_thrs
Ps = fg_w_thrs + bg_w_thrs
Ps[Ps == 0] = 1
T = max(np.count_nonzero(gt), 1)
precisions = TPs / Ps
recalls = TPs / T
numerator = (1 + self.beta) * precisions * recalls
denominator = np.where(numerator == 0, 1, self.beta * precisions + recalls)
changeable_fms = numerator / denominator
return precisions, recalls, changeable_fms
def get_results(self):
adaptive_fm = np.mean(np.array(self.adaptive_fms, _TYPE))
# precision = np.mean(np.array(self.precisions, dtype=_TYPE), axis=0) # N, 256
# recall = np.mean(np.array(self.recalls, dtype=_TYPE), axis=0) # N, 256
changeable_fm = np.mean(np.array(self.changeable_fms, dtype=_TYPE), axis=0)
# return dict(fm=dict(adp=adaptive_fm, curve=changeable_fm),
# pr=dict(p=precision, r=recall))
return dict(adpFm=adaptive_fm, meanFm=changeable_fm, maxFm=changeable_fm)
class MAE(object):
def __init__(self, length):
self.maes = []
def step(self, pred: np.ndarray, gt: np.ndarray, idx):
pred, gt = _prepare_data(pred, gt)
mae = self.cal_mae(pred, gt)
self.maes.append(mae)
def cal_mae(self, pred: np.ndarray, gt: np.ndarray) -> float:
mae = np.mean(np.abs(pred - gt))
return mae
def get_results(self):
mae = np.mean(np.array(self.maes, _TYPE))
return dict(MAE=mae)
class Smeasure(object):
def __init__(self, length, alpha: float = 0.5):
self.sms = []
self.alpha = alpha
def step(self, pred: np.ndarray, gt: np.ndarray, idx):
pred, gt = _prepare_data(pred=pred, gt=gt)
sm = self.cal_sm(pred, gt)
self.sms.append(sm)
def cal_sm(self, pred: np.ndarray, gt: np.ndarray) -> float:
y = np.mean(gt)
if y == 0:
sm = 1 - np.mean(pred)
elif y == 1:
sm = np.mean(pred)
else:
sm = self.alpha * self.object(pred, gt) + (1 - self.alpha) * self.region(pred, gt)
sm = max(0, sm)
return sm
def object(self, pred: np.ndarray, gt: np.ndarray) -> float:
fg = pred * gt
bg = (1 - pred) * (1 - gt)
u = np.mean(gt)
object_score = u * self.s_object(fg, gt) + (1 - u) * self.s_object(bg, 1 - gt)
return object_score
def s_object(self, pred: np.ndarray, gt: np.ndarray) -> float:
x = np.mean(pred[gt == 1])
sigma_x = np.std(pred[gt == 1], ddof=1)
score = 2 * x / (np.power(x, 2) + 1 + sigma_x + _EPS)
return score
def region(self, pred: np.ndarray, gt: np.ndarray) -> float:
x, y = self.centroid(gt)
part_info = self.divide_with_xy(pred, gt, x, y)
w1, w2, w3, w4 = part_info['weight']
pred1, pred2, pred3, pred4 = part_info['pred']
gt1, gt2, gt3, gt4 = part_info['gt']
score1 = self.ssim(pred1, gt1)
score2 = self.ssim(pred2, gt2)
score3 = self.ssim(pred3, gt3)
score4 = self.ssim(pred4, gt4)
return w1 * score1 + w2 * score2 + w3 * score3 + w4 * score4
def centroid(self, matrix: np.ndarray) -> tuple:
"""
To ensure consistency with the matlab code, one is added to the centroid coordinate,
so there is no need to use the redundant addition operation when dividing the region later,
because the sequence generated by ``1:X`` in matlab will contain ``X``.
:param matrix: a bool data array
:return: the centroid coordinate
"""
h, w = matrix.shape
area_object = np.count_nonzero(matrix)
if area_object == 0:
x = np.round(w / 2)
y = np.round(h / 2)
else:
# More details can be found at: https://www.yuque.com/lart/blog/gpbigm
y, x = np.argwhere(matrix).mean(axis=0).round()
return int(x) + 1, int(y) + 1
def divide_with_xy(self, pred: np.ndarray, gt: np.ndarray, x, y) -> dict:
h, w = gt.shape
area = h * w
gt_LT = gt[0:y, 0:x]
gt_RT = gt[0:y, x:w]
gt_LB = gt[y:h, 0:x]
gt_RB = gt[y:h, x:w]
pred_LT = pred[0:y, 0:x]
pred_RT = pred[0:y, x:w]
pred_LB = pred[y:h, 0:x]
pred_RB = pred[y:h, x:w]
w1 = x * y / area
w2 = y * (w - x) / area
w3 = (h - y) * x / area
w4 = 1 - w1 - w2 - w3
return dict(gt=(gt_LT, gt_RT, gt_LB, gt_RB),
pred=(pred_LT, pred_RT, pred_LB, pred_RB),
weight=(w1, w2, w3, w4))
def ssim(self, pred: np.ndarray, gt: np.ndarray) -> float:
h, w = pred.shape
N = h * w
x = np.mean(pred)
y = np.mean(gt)
sigma_x = np.sum((pred - x) ** 2) / (N - 1)
sigma_y = np.sum((gt - y) ** 2) / (N - 1)
sigma_xy = np.sum((pred - x) * (gt - y)) / (N - 1)
alpha = 4 * x * y * sigma_xy
beta = (x ** 2 + y ** 2) * (sigma_x + sigma_y)
if alpha != 0:
score = alpha / (beta + _EPS)
elif alpha == 0 and beta == 0:
score = 1
else:
score = 0
return score
def get_results(self):
sm = np.mean(np.array(self.sms, dtype=_TYPE))
return dict(Smeasure=sm)
class Emeasure(object):
def __init__(self, length):
self.adaptive_ems = []
self.changeable_ems = []
def step(self, pred: np.ndarray, gt: np.ndarray, idx):
pred, gt = _prepare_data(pred=pred, gt=gt)
self.gt_fg_numel = np.count_nonzero(gt)
self.gt_size = gt.shape[0] * gt.shape[1]
changeable_ems = self.cal_changeable_em(pred, gt)
self.changeable_ems.append(changeable_ems)
adaptive_em = self.cal_adaptive_em(pred, gt)
self.adaptive_ems.append(adaptive_em)
def cal_adaptive_em(self, pred: np.ndarray, gt: np.ndarray) -> float:
adaptive_threshold = _get_adaptive_threshold(pred, max_value=1)
adaptive_em = self.cal_em_with_threshold(pred, gt, threshold=adaptive_threshold)
return adaptive_em
def cal_changeable_em(self, pred: np.ndarray, gt: np.ndarray) -> np.ndarray:
changeable_ems = self.cal_em_with_cumsumhistogram(pred, gt)
return changeable_ems
def cal_em_with_threshold(self, pred: np.ndarray, gt: np.ndarray, threshold: float) -> float:
binarized_pred = pred >= threshold
fg_fg_numel = np.count_nonzero(binarized_pred & gt)
fg_bg_numel = np.count_nonzero(binarized_pred & ~gt)
fg___numel = fg_fg_numel + fg_bg_numel
bg___numel = self.gt_size - fg___numel
if self.gt_fg_numel == 0:
enhanced_matrix_sum = bg___numel
elif self.gt_fg_numel == self.gt_size:
enhanced_matrix_sum = fg___numel
else:
parts_numel, combinations = self.generate_parts_numel_combinations(
fg_fg_numel=fg_fg_numel, fg_bg_numel=fg_bg_numel,
pred_fg_numel=fg___numel, pred_bg_numel=bg___numel,
)
results_parts = []
for i, (part_numel, combination) in enumerate(zip(parts_numel, combinations)):
align_matrix_value = 2 * (combination[0] * combination[1]) / \
(combination[0] ** 2 + combination[1] ** 2 + _EPS)
enhanced_matrix_value = (align_matrix_value + 1) ** 2 / 4
results_parts.append(enhanced_matrix_value * part_numel)
enhanced_matrix_sum = sum(results_parts)
em = enhanced_matrix_sum / (self.gt_size - 1 + _EPS)
return em
def cal_em_with_cumsumhistogram(self, pred: np.ndarray, gt: np.ndarray) -> np.ndarray:
pred = (pred * 255).astype(np.uint8)
bins = np.linspace(0, 256, 257)
fg_fg_hist, _ = np.histogram(pred[gt], bins=bins)
fg_bg_hist, _ = np.histogram(pred[~gt], bins=bins)
fg_fg_numel_w_thrs = np.cumsum(np.flip(fg_fg_hist), axis=0)
fg_bg_numel_w_thrs = np.cumsum(np.flip(fg_bg_hist), axis=0)
fg___numel_w_thrs = fg_fg_numel_w_thrs + fg_bg_numel_w_thrs
bg___numel_w_thrs = self.gt_size - fg___numel_w_thrs
if self.gt_fg_numel == 0:
enhanced_matrix_sum = bg___numel_w_thrs
elif self.gt_fg_numel == self.gt_size:
enhanced_matrix_sum = fg___numel_w_thrs
else:
parts_numel_w_thrs, combinations = self.generate_parts_numel_combinations(
fg_fg_numel=fg_fg_numel_w_thrs, fg_bg_numel=fg_bg_numel_w_thrs,
pred_fg_numel=fg___numel_w_thrs, pred_bg_numel=bg___numel_w_thrs,
)
results_parts = np.empty(shape=(4, 256), dtype=np.float64)
for i, (part_numel, combination) in enumerate(zip(parts_numel_w_thrs, combinations)):
align_matrix_value = 2 * (combination[0] * combination[1]) / \
(combination[0] ** 2 + combination[1] ** 2 + _EPS)
enhanced_matrix_value = (align_matrix_value + 1) ** 2 / 4
results_parts[i] = enhanced_matrix_value * part_numel
enhanced_matrix_sum = results_parts.sum(axis=0)
em = enhanced_matrix_sum / (self.gt_size - 1 + _EPS)
return em
def generate_parts_numel_combinations(self, fg_fg_numel, fg_bg_numel, pred_fg_numel, pred_bg_numel):
bg_fg_numel = self.gt_fg_numel - fg_fg_numel
bg_bg_numel = pred_bg_numel - bg_fg_numel
parts_numel = [fg_fg_numel, fg_bg_numel, bg_fg_numel, bg_bg_numel]
mean_pred_value = pred_fg_numel / self.gt_size
mean_gt_value = self.gt_fg_numel / self.gt_size
demeaned_pred_fg_value = 1 - mean_pred_value
demeaned_pred_bg_value = 0 - mean_pred_value
demeaned_gt_fg_value = 1 - mean_gt_value
demeaned_gt_bg_value = 0 - mean_gt_value
combinations = [
(demeaned_pred_fg_value, demeaned_gt_fg_value),
(demeaned_pred_fg_value, demeaned_gt_bg_value),
(demeaned_pred_bg_value, demeaned_gt_fg_value),
(demeaned_pred_bg_value, demeaned_gt_bg_value)
]
return parts_numel, combinations
def get_results(self):
adaptive_em = np.mean(np.array(self.adaptive_ems, dtype=_TYPE))
changeable_em = np.mean(np.array(self.changeable_ems, dtype=_TYPE), axis=0)
return dict(adpEm=adaptive_em, meanEm=changeable_em, maxEm=changeable_em)
class WeightedFmeasure(object):
def __init__(self, length, beta: float = 1):
self.beta = beta
self.weighted_fms = []
def step(self, pred: np.ndarray, gt: np.ndarray, idx):
pred, gt = _prepare_data(pred=pred, gt=gt)
if np.all(~gt):
wfm = 0
else:
wfm = self.cal_wfm(pred, gt)
self.weighted_fms.append(wfm)
def cal_wfm(self, pred: np.ndarray, gt: np.ndarray) -> float:
# [Dst,IDXT] = bwdist(dGT);
Dst, Idxt = bwdist(gt == 0, return_indices=True)
# %Pixel dependency
# E = abs(FG-dGT);
E = np.abs(pred - gt)
Et = np.copy(E)
Et[gt == 0] = Et[Idxt[0][gt == 0], Idxt[1][gt == 0]]
# K = fspecial('gaussian',7,5);
# EA = imfilter(Et,K);
K = self.matlab_style_gauss2D((7, 7), sigma=5)
EA = convolve(Et, weights=K, mode="constant", cval=0)
# MIN_E_EA = E;
# MIN_E_EA(GT & EA<E) = EA(GT & EA<E);
MIN_E_EA = np.where(gt & (EA < E), EA, E)
# %Pixel importance
B = np.where(gt == 0, 2 - np.exp(np.log(0.5) / 5 * Dst), np.ones_like(gt))
Ew = MIN_E_EA * B
TPw = np.sum(gt) - np.sum(Ew[gt == 1])
FPw = np.sum(Ew[gt == 0])
R = 1 - np.mean(Ew[gt == 1])
P = TPw / (TPw + FPw + _EPS)
# % Q = (1+Beta^2)*(R*P)./(eps+R+(Beta.*P));
Q = (1 + self.beta) * R * P / (R + self.beta * P + _EPS)
return Q
def matlab_style_gauss2D(self, shape: tuple = (7, 7), sigma: int = 5) -> np.ndarray:
"""
2D gaussian mask - should give the same result as MATLAB's
fspecial('gaussian',[shape],[sigma])
"""
m, n = [(ss - 1) / 2 for ss in shape]
y, x = np.ogrid[-m: m + 1, -n: n + 1]
h = np.exp(-(x * x + y * y) / (2 * sigma * sigma))
h[h < np.finfo(h.dtype).eps * h.max()] = 0
sumh = h.sum()
if sumh != 0:
h /= sumh
return h
def get_results(self):
weighted_fm = np.mean(np.array(self.weighted_fms, dtype=_TYPE))
return dict(wFmeasure=weighted_fm)
class Medical(object):
def __init__(self, length):
self.Thresholds = np.linspace(1, 0, 256)
self.threshold_Sensitivity = np.zeros((length, len(self.Thresholds)))
self.threshold_Specificity = np.zeros((length, len(self.Thresholds)))
self.threshold_Dice = np.zeros((length, len(self.Thresholds)))
self.threshold_IoU = np.zeros((length, len(self.Thresholds)))
def Fmeasure_calu(self, pred, gt, threshold):
if threshold > 1:
threshold = 1
Label3 = np.zeros_like(gt)
Label3[pred >= threshold] = 1
NumRec = np.sum(Label3 == 1)
NumNoRec = np.sum(Label3 == 0)
LabelAnd = (Label3 == 1) & (gt == 1)
NumAnd = np.sum(LabelAnd == 1)
num_obj = np.sum(gt)
num_pred = np.sum(Label3)
FN = num_obj - NumAnd
FP = NumRec - NumAnd
TN = NumNoRec - FN
if NumAnd == 0:
RecallFtem = 0
Dice = 0
SpecifTem = 0
IoU = 0
else:
IoU = NumAnd / (FN + NumRec)
RecallFtem = NumAnd / num_obj
SpecifTem = TN / (TN + FP)
Dice = 2 * NumAnd / (num_obj + num_pred)
return RecallFtem, SpecifTem, Dice, IoU
def step(self, pred, gt, idx):
pred, gt = _prepare_data(pred=pred, gt=gt)
threshold_Rec = np.zeros(len(self.Thresholds))
threshold_Iou = np.zeros(len(self.Thresholds))
threshold_Spe = np.zeros(len(self.Thresholds))
threshold_Dic = np.zeros(len(self.Thresholds))
for j, threshold in enumerate(self.Thresholds):
threshold_Rec[j], threshold_Spe[j], threshold_Dic[j], \
threshold_Iou[j] = self.Fmeasure_calu(pred, gt, threshold)
self.threshold_Sensitivity[idx, :] = threshold_Rec
self.threshold_Specificity[idx, :] = threshold_Spe
self.threshold_Dice[idx, :] = threshold_Dic
self.threshold_IoU[idx, :] = threshold_Iou
def get_results(self):
column_Sen = np.mean(self.threshold_Sensitivity, axis=0)
column_Spe = np.mean(self.threshold_Specificity, axis=0)
column_Dic = np.mean(self.threshold_Dice, axis=0)
column_IoU = np.mean(self.threshold_IoU, axis=0)
return dict(meanSen=column_Sen, meanSpe=column_Spe, meanDice=column_Dic, meanIoU=column_IoU,
maxSen=column_Sen, maxSpe=column_Spe, maxDice=column_Dic, maxIoU=column_IoU)
================================================
FILE: eval/vps_evaluator.py
================================================
# -*- coding: utf-8 -*-
# @Time : 2022/03/14
# @Author : Johnson-Chou
# @Email : johnson111788@gmail.com
# @FileName : vps_evaluator.py
import glob
import os
import cv2
import argparse
from tqdm import tqdm
import prettytable as pt
import numpy as np
def get_competitors(root):
for model_name in os.listdir(root):
print('\'{}\''.format(model_name), end=', ')
def evaluator(gt_pth_lst, pred_pth_lst, metrics):
module_map_name = {"Smeasure": "Smeasure", "wFmeasure": "WeightedFmeasure", "MAE": "MAE",
"adpEm": "Emeasure", "meanEm": "Emeasure", "maxEm": "Emeasure",
"adpFm": "Fmeasure", "meanFm": "Fmeasure", "maxFm": "Fmeasure",
"meanSen": "Medical", "maxSen": "Medical", "meanSpe": "Medical", "maxSpe": "Medical",
"meanDice": "Medical", "maxDice": "Medical", "meanIoU": "Medical", "maxIoU": "Medical"}
res, metric_module = {}, {}
metric_module_list = [module_map_name[metric] for metric in metrics]
metric_module_list = list(set(metric_module_list))
# define measures
for metric_module_name in metric_module_list:
metric_module[metric_module_name] = getattr(__import__("metrics", fromlist=[metric_module_name]),
metric_module_name)(length=len(gt_pth_lst))
assert len(gt_pth_lst) == len(pred_pth_lst)
# evaluator
for idx in tqdm(range(len(gt_pth_lst))):
gt_pth = gt_pth_lst[idx]
pred_pth = pred_pth_lst[idx]
# print(gt_pth, pred_pth)
assert os.path.isfile(gt_pth) and os.path.isfile(pred_pth)
pred_ary = cv2.imread(pred_pth, cv2.IMREAD_GRAYSCALE)
gt_ary = cv2.imread(gt_pth, cv2.IMREAD_GRAYSCALE)
# ensure the shape of prediction is matched to gt
if not gt_ary.shape == pred_ary.shape:
pred_ary = cv2.resize(pred_ary, (gt_ary.shape[1], gt_ary.shape[0]))
for module in metric_module.values():
module.step(pred=pred_ary, gt=gt_ary, idx=idx)
for metric in metrics:
module = metric_module[module_map_name[metric]]
res[metric] = module.get_results()[metric]
return res
def eval_engine_vps(opt, txt_save_path):
# evaluation for whole dataset
for _data_name in opt.data_lst[0]:
print('#' * 20, 'Current Dataset:', _data_name, '#' * 20)
filename = os.path.join(txt_save_path, '{}_eval.txt'.format(_data_name.replace('/', '-')))
with open(filename, 'w+') as file_to_write:
# initial settings for PrettyTable
tb = pt.PrettyTable()
names = ["Dataset", "Method"]
names.extend(opt.metric_list)
tb.field_names = names
# iter each method for current dataset
for _model_name in opt.model_lst[0]:
print('#' * 10, 'Current Method:', _model_name, '#' * 10)
gt_src = os.path.join(opt.gt_root, _data_name, 'GT')
pred_src = os.path.join(opt.pred_root, _model_name, _data_name)
# get the sequence list for current dataset
case_list = os.listdir(gt_src)
mean_case_score_list, max_case_score_list = [], []
# iter each video frame for current method-dataset
for case in case_list:
case_gt_name_list = glob.glob(gt_src + '/{}/*.png'.format(case))
try:
case_gt_name_list.sort(
key=lambda name: (int(name.split('/')[-2]), int(name.split('/')[-1].rstrip('.png')))
)
except:
case_gt_name_list.sort(
key=lambda name: (int(name.split("/")[-2].split('case')[1].split('_')[0]),
0 if not len(name.split('/')[-2].split('_')) > 1 else int(
name.split('/')[-2].split('_')[1]),
int(name.split('/')[-1].split('-')[0].split('_')[-1]),
int(name.split('/')[-1].split('_a')[1].split('_')[0]),
int(name.split('/')[-1].split('_image')[1].split('.png')[
0])))
# for fair comparison, we remove the first frame and last frame in the video suggested by reference: Shifting More Attention to Video Salient Object Detection
# https://github.com/DengPingFan/DAVSOD/blob/master/EvaluateTool/main.m
case_gt_name_list = case_gt_name_list[1:-1]
case_pred_name_list = [gt.replace(gt_src, pred_src) for gt in case_gt_name_list]
result = evaluator(
gt_pth_lst=case_gt_name_list,
pred_pth_lst=case_pred_name_list,
metrics=opt.metric_list
)
mean_score_ind, max_score_ind = [], []
mean_score_list, max_score_list = [], []
for i, (name, value) in enumerate(result.items()):
if 'max' in name or 'mean' in name:
if 'max' in name:
max_score_list.append(value)
max_score_ind.append(i)
else:
mean_score_list.append(value)
mean_score_ind.append(i)
else:
mean_score_list.append([value]*256)
mean_score_ind.append(i)
# calculate all the metrics at frame-level
max_case_score_list.append(max_score_list)
mean_case_score_list.append(mean_score_list)
# calculate all the metrics at sequence-level
max_case_score_list = np.mean(np.array(max_case_score_list), axis=0)
mean_case_score_list = np.mean(np.array(mean_case_score_list), axis=0)
case_score_list = []
for index in range(len(opt.metric_list)):
real_max_index = np.where(np.array(max_score_ind) == index)
real_mean_index = np.where(np.array(mean_score_ind) == index)
if len(real_max_index[0]) > 0:
case_score_list.append(max_case_score_list[real_max_index[0]].max().round(3))
else:
case_score_list.append(mean_case_score_list[real_mean_index[0]].mean().round(3))
final_score_list = ['{:.3f}'.format(case) for case in case_score_list]
tb.add_row([_data_name.replace('/', '-'), _model_name] + list(final_score_list))
print(tb)
file_to_write.write(str(tb))
file_to_write.close()
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument(
'--gt_root', type=str, help='custom your ground-truth root',
default='../data/SUN-SEG-Annotation/')
parser.add_argument(
'--pred_root', type=str, help='custom your prediction root',
default='../data/Pred/')
parser.add_argument(
'--metric_list', type=list, help='set the evaluation metrics',
default=['Smeasure', 'maxEm', 'wFmeasure', 'maxDice', 'maxIoU'],
choices=["Smeasure", "wFmeasure", "MAE", "adpEm", "meanEm", "maxEm", "adpFm", "meanFm", "maxFm",
"meanSen", "maxSen", "meanSpe", "maxSpe", "meanDice", "maxDice", "meanIoU", "maxIoU"])
parser.add_argument(
'--data_lst', type=str, help='set the dataset what you wanna to test',
nargs='+', action='append',
choices=['TestEasyDataset/Seen', 'TestHardDataset/Seen', 'TestEasyDataset/Unseen', 'TestHardDataset/Unseen'])
parser.add_argument(
'--model_lst', type=str, help='candidate competitors',
gitextract_qn90sj5e/
├── .gitignore
├── LICENSE
├── README.md
├── docs/
│ ├── AWESOME_VPS.md
│ ├── DATA_DESCRIPTION.md
│ ├── DATA_PREPARATION.md
│ ├── INFO_NEGATIVE_CASES.md
│ ├── INFO_POSITIVE_CASES.md
│ └── RELEASE_NOTES.md
├── eval/
│ ├── README.md
│ ├── eval-result/
│ │ ├── 2015-MICCAI-UNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2018-TMI-UNet++/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2019-TPAMI-COSNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2020-AAAI-PCSA/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2020-MICCAI-23DCNN/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2020-MICCAI-ACSNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2020-MICCAI-PraNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2020-TIP-MATNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2021-ICCV-DCFNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2021-ICCV-FSNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2021-MICCAI-PNSNet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2021-MICCAI-SANet/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ ├── 2021-NIPS-AMD/
│ │ │ ├── TestEasyDataset_eval.txt
│ │ │ └── TestHardDataset_eval.txt
│ │ └── 2022-TMI-PNSPlus/
│ │ ├── TestEasyDataset_eval.txt
│ │ └── TestHardDataset_eval.txt
│ ├── eval.sh
│ ├── metrics.py
│ └── vps_evaluator.py
├── lib/
│ ├── __init__.py
│ ├── dataloader/
│ │ ├── __init__.py
│ │ ├── dataloader.py
│ │ ├── preprocess.py
│ │ └── statistics.pth
│ ├── module/
│ │ ├── LightRFB.py
│ │ ├── PNS/
│ │ │ ├── PNS_Module/
│ │ │ │ ├── CMakeLists.txt
│ │ │ │ ├── reference.cpp
│ │ │ │ ├── reference.h
│ │ │ │ ├── sa.cu
│ │ │ │ ├── sa_ext.cpp
│ │ │ │ ├── timer.h
│ │ │ │ └── utils.h
│ │ │ └── setup.py
│ │ ├── PNSPlusModule.py
│ │ ├── PNSPlusNetwork.py
│ │ ├── Res2Net_v1b.py
│ │ └── __init__.py
│ └── utils/
│ ├── __init__.py
│ └── utils.py
├── scripts/
│ ├── config.py
│ ├── eval_eff.py
│ ├── my_test.py
│ └── my_train.py
├── snapshot/
│ └── .placeholder
└── utils/
├── reorganize.py
└── reorganize.sh
SYMBOL INDEX (152 symbols across 15 files)
FILE: eval/metrics.py
function _prepare_data (line 17) | def _prepare_data(pred: np.ndarray, gt: np.ndarray) -> tuple:
function _get_adaptive_threshold (line 25) | def _get_adaptive_threshold(matrix: np.ndarray, max_value: float = 1) ->...
class Fmeasure (line 29) | class Fmeasure(object):
method __init__ (line 30) | def __init__(self, length, beta: float = 0.3):
method step (line 37) | def step(self, pred: np.ndarray, gt: np.ndarray, idx):
method cal_adaptive_fm (line 48) | def cal_adaptive_fm(self, pred: np.ndarray, gt: np.ndarray) -> float:
method cal_pr (line 61) | def cal_pr(self, pred: np.ndarray, gt: np.ndarray) -> tuple:
method get_results (line 79) | def get_results(self):
class MAE (line 89) | class MAE(object):
method __init__ (line 90) | def __init__(self, length):
method step (line 93) | def step(self, pred: np.ndarray, gt: np.ndarray, idx):
method cal_mae (line 99) | def cal_mae(self, pred: np.ndarray, gt: np.ndarray) -> float:
method get_results (line 103) | def get_results(self):
class Smeasure (line 108) | class Smeasure(object):
method __init__ (line 109) | def __init__(self, length, alpha: float = 0.5):
method step (line 113) | def step(self, pred: np.ndarray, gt: np.ndarray, idx):
method cal_sm (line 119) | def cal_sm(self, pred: np.ndarray, gt: np.ndarray) -> float:
method object (line 130) | def object(self, pred: np.ndarray, gt: np.ndarray) -> float:
method s_object (line 137) | def s_object(self, pred: np.ndarray, gt: np.ndarray) -> float:
method region (line 143) | def region(self, pred: np.ndarray, gt: np.ndarray) -> float:
method centroid (line 156) | def centroid(self, matrix: np.ndarray) -> tuple:
method divide_with_xy (line 174) | def divide_with_xy(self, pred: np.ndarray, gt: np.ndarray, x, y) -> dict:
method ssim (line 197) | def ssim(self, pred: np.ndarray, gt: np.ndarray) -> float:
method get_results (line 219) | def get_results(self):
class Emeasure (line 224) | class Emeasure(object):
method __init__ (line 225) | def __init__(self, length):
method step (line 229) | def step(self, pred: np.ndarray, gt: np.ndarray, idx):
method cal_adaptive_em (line 239) | def cal_adaptive_em(self, pred: np.ndarray, gt: np.ndarray) -> float:
method cal_changeable_em (line 244) | def cal_changeable_em(self, pred: np.ndarray, gt: np.ndarray) -> np.nd...
method cal_em_with_threshold (line 248) | def cal_em_with_threshold(self, pred: np.ndarray, gt: np.ndarray, thre...
method cal_em_with_cumsumhistogram (line 277) | def cal_em_with_cumsumhistogram(self, pred: np.ndarray, gt: np.ndarray...
method generate_parts_numel_combinations (line 309) | def generate_parts_numel_combinations(self, fg_fg_numel, fg_bg_numel, ...
method get_results (line 331) | def get_results(self):
class WeightedFmeasure (line 337) | class WeightedFmeasure(object):
method __init__ (line 338) | def __init__(self, length, beta: float = 1):
method step (line 342) | def step(self, pred: np.ndarray, gt: np.ndarray, idx):
method cal_wfm (line 351) | def cal_wfm(self, pred: np.ndarray, gt: np.ndarray) -> float:
method matlab_style_gauss2D (line 385) | def matlab_style_gauss2D(self, shape: tuple = (7, 7), sigma: int = 5) ...
method get_results (line 399) | def get_results(self):
class Medical (line 404) | class Medical(object):
method __init__ (line 405) | def __init__(self, length):
method Fmeasure_calu (line 413) | def Fmeasure_calu(self, pred, gt, threshold):
method step (line 446) | def step(self, pred, gt, idx):
method get_results (line 463) | def get_results(self):
FILE: eval/vps_evaluator.py
function get_competitors (line 15) | def get_competitors(root):
function evaluator (line 20) | def evaluator(gt_pth_lst, pred_pth_lst, metrics):
function eval_engine_vps (line 61) | def eval_engine_vps(opt, txt_save_path):
FILE: lib/dataloader/dataloader.py
class VideoDataset (line 10) | class VideoDataset(Dataset):
method __init__ (line 11) | def __init__(self, video_dataset, transform=None, time_interval=1):
method __getitem__ (line 50) | def __getitem__(self, idx):
method __len__ (line 74) | def __len__(self):
function get_video_dataset (line 78) | def get_video_dataset():
FILE: lib/dataloader/preprocess.py
class Compose_imglabel (line 7) | class Compose_imglabel(object):
method __init__ (line 8) | def __init__(self, transforms):
method __call__ (line 11) | def __call__(self, img, label):
class Random_crop_Resize_Video (line 17) | class Random_crop_Resize_Video(object):
method _randomCrop (line 18) | def _randomCrop(self, img, label, x, y):
method __init__ (line 26) | def __init__(self, crop_size):
method __call__ (line 29) | def __call__(self, imgs, labels):
class Random_horizontal_flip_video (line 40) | class Random_horizontal_flip_video(object):
method _horizontal_flip (line 41) | def _horizontal_flip(self, img, label):
method __init__ (line 44) | def __init__(self, prob):
method __call__ (line 51) | def __call__(self, imgs, labels):
class Resize_video (line 70) | class Resize_video(object):
method __init__ (line 71) | def __init__(self, height, width):
method __call__ (line 75) | def __call__(self, imgs, labels):
class Normalize_video (line 84) | class Normalize_video(object):
method __init__ (line 85) | def __init__(self, mean, std):
method __call__ (line 88) | def __call__(self, imgs, labels):
class toTensor_video (line 99) | class toTensor_video(object):
method __init__ (line 100) | def __init__(self):
method __call__ (line 103) | def __call__(self, imgs, labels):
FILE: lib/module/LightRFB.py
class h_sigmoid (line 5) | class h_sigmoid(nn.Module):
method __init__ (line 6) | def __init__(self, inplace=True):
method forward (line 10) | def forward(self, x):
class SELayer (line 14) | class SELayer(nn.Module):
method __init__ (line 15) | def __init__(self, channel, reduction=4):
method forward (line 25) | def forward(self, x):
class BasicConv (line 32) | class BasicConv(nn.Module):
method __init__ (line 34) | def __init__(self, in_planes, out_planes, kernel_size, stride=1, paddi...
method forward (line 43) | def forward(self, x):
class LightRFB (line 52) | class LightRFB(nn.Module):
method __init__ (line 53) | def __init__(self, channels_in=1024, channels_mid=128, channels_out=32):
method forward (line 92) | def forward(self, x):
FILE: lib/module/PNS/PNS_Module/reference.cpp
function sa_weight_forward_Ref (line 7) | void sa_weight_forward_Ref(const torch::Tensor& query,const torch::Tenso...
function sa_weight_backward_query_Ref (line 37) | void sa_weight_backward_query_Ref(const torch::Tensor& dweight,const tor...
function sa_weight_backward_key_Ref (line 68) | void sa_weight_backward_key_Ref(const torch::Tensor& dweight,const torch...
function sa_map_forward_Ref (line 95) | void sa_map_forward_Ref(const torch::Tensor& weight,const torch::Tensor&...
function sa_map_backward_weight_Ref (line 125) | void sa_map_backward_weight_Ref(const torch::Tensor& dout,const torch::T...
function sa_map_backward_proj_Ref (line 156) | void sa_map_backward_proj_Ref(const torch::Tensor& dout,const torch::Ten...
FILE: lib/module/PNS/PNS_Module/sa_ext.cpp
function get_sizes (line 7) | void get_sizes(const torch::Tensor& t,int *B,int *T,int *C,int *H,int *W){
function sa_weight_forward (line 26) | void sa_weight_forward(const torch::Tensor& query,const torch::Tensor& k...
function sa_map_forward (line 38) | void sa_map_forward(const torch::Tensor& weight,const torch::Tensor& pro...
function sa_weight_backward (line 50) | void sa_weight_backward(const torch::Tensor& dw,const torch::Tensor& que...
function sa_map_backward (line 62) | void sa_map_backward(const torch::Tensor& dout,const torch::Tensor& weig...
function sa_weight_forward_ref (line 74) | void sa_weight_forward_ref(const torch::Tensor& query,const torch::Tenso...
function sa_weight_backward_ref (line 80) | void sa_weight_backward_ref(const torch::Tensor& dw,const torch::Tensor&...
function sa_map_forward_ref (line 87) | void sa_map_forward_ref(const torch::Tensor& weight,const torch::Tensor&...
function sa_map_backward_ref (line 93) | void sa_map_backward_ref(const torch::Tensor& dout,const torch::Tensor& ...
function PYBIND11_MODULE (line 101) | PYBIND11_MODULE(TORCH_EXTENSION_NAME, m) {
FILE: lib/module/PNS/PNS_Module/timer.h
function Start (line 23) | void Start()
function Stop (line 28) | void Stop()
function Elapsed (line 33) | float Elapsed()
FILE: lib/module/PNSPlusModule.py
function _check_contiguous (line 13) | def _check_contiguous(*args):
class Relevance_Measuring (line 18) | class Relevance_Measuring(autograd.Function):
method forward (line 20) | def forward(ctx, query, key, radius=1, dilation=1):
method backward (line 36) | def backward(ctx, dw):
class Spatial_Temporal_Aggregation (line 45) | class Spatial_Temporal_Aggregation(autograd.Function):
method forward (line 47) | def forward(ctx, weight, proj, radius=1, dilation=1):
method backward (line 58) | def backward(ctx, dout):
class NS_Block (line 71) | class NS_Block(nn.Module):
method __init__ (line 72) | def __init__(self, channels_in=32, n_head=4, d_k=8, d_v=8, radius=[3, ...
method forward (line 86) | def forward(self, first, x):
FILE: lib/module/PNSPlusNetwork.py
class conbine_feature (line 12) | class conbine_feature(nn.Module):
method __init__ (line 13) | def __init__(self):
method forward (line 21) | def forward(self, low_fea, high_fea):
class DilatedParallelConvBlockD2 (line 28) | class DilatedParallelConvBlockD2(nn.Module):
method __init__ (line 29) | def __init__(self, nIn, nOut, add=False):
method forward (line 41) | def forward(self, input):
class PNSNet (line 55) | class PNSNet(nn.Module):
method __init__ (line 56) | def __init__(self):
method forward (line 68) | def forward(self, x):
FILE: lib/module/Res2Net_v1b.py
class Bottle2neck (line 15) | class Bottle2neck(nn.Module):
method __init__ (line 18) | def __init__(self, inplanes, planes, stride=1, downsample=None, baseWi...
method forward (line 58) | def forward(self, x):
class Res2Net (line 94) | class Res2Net(nn.Module):
method __init__ (line 96) | def __init__(self, block, layers, baseWidth=26, scale=4, num_classes=1...
method _make_layer (line 127) | def _make_layer(self, block, planes, blocks, stride=1):
method forward (line 147) | def forward(self, x):
function res2net50_v1b (line 165) | def res2net50_v1b(pretrained=False, **kwargs):
function res2net101_v1b (line 177) | def res2net101_v1b(pretrained=False, **kwargs):
function res2net50_v1b_26w_4s (line 188) | def res2net50_v1b_26w_4s(pretrained=False, **kwargs):
function res2net101_v1b_26w_4s (line 202) | def res2net101_v1b_26w_4s(pretrained=False, **kwargs):
function res2net152_v1b_26w_4s (line 213) | def res2net152_v1b_26w_4s(pretrained=False, **kwargs):
FILE: lib/utils/utils.py
function clip_gradient (line 1) | def clip_gradient(optimizer, grad_clip):
function adjust_lr (line 14) | def adjust_lr(optimizer, init_lr, epoch, decay_rate=0.1, decay_epoch=30):
FILE: scripts/eval_eff.py
function computeTime (line 11) | def computeTime(model, inputs, device='cuda'):
FILE: scripts/my_test.py
function safe_save (line 14) | def safe_save(img, save_path):
class Normalize (line 19) | class Normalize(object):
method __init__ (line 20) | def __init__(self, mean, std):
method __call__ (line 23) | def __call__(self, img):
class Test_Dataset (line 31) | class Test_Dataset(Dataset):
method __init__ (line 32) | def __init__(self, root, testset):
method __getitem__ (line 76) | def __getitem__(self, idx):
method __len__ (line 91) | def __len__(self):
class AutoTest (line 95) | class AutoTest:
method __init__ (line 96) | def __init__(self, test_dataset, data_root, model_path):
method test (line 114) | def test(self):
FILE: scripts/my_train.py
class CrossEntropyLoss (line 17) | class CrossEntropyLoss(nn.Module):
method __init__ (line 18) | def __init__(self):
method forward (line 21) | def forward(self, *inputs):
function train (line 27) | def train(train_loader, model, optimizer, epoch, save_path, loss_func):
Condensed preview — 68 files, each showing path, character count, and a content snippet. Download the .json file or copy for the full structured content (282K chars).
[
{
"path": ".gitignore",
"chars": 269,
"preview": "# For projects using Nanoc (http://nanoc.ws/)\n\n# Default location for output (needs to match output_dir's value found in"
},
{
"path": "LICENSE",
"chars": 11357,
"preview": " Apache License\n Version 2.0, January 2004\n "
},
{
"path": "README.md",
"chars": 17002,
"preview": "# <p align=center>`Video Polyp Segmentation: A Deep Learning Perspective (MIR 2022)`</p><!-- omit in toc -->\n\n[ for the feedback and for fixing the sorting bug in ou"
},
{
"path": "eval/README.md",
"chars": 2299,
"preview": "# VPS Evaluation Toolbox\n\nThis toolbox is used to evaluate the performance of video polyp segmentation task.\n\n# Usage\n\n-"
},
{
"path": "eval/eval-result/2015-MICCAI-UNet/TestEasyDataset_eval.txt",
"chars": 484,
"preview": "+-----------------+------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2015-MICCAI-UNet/TestHardDataset_eval.txt",
"chars": 484,
"preview": "+-----------------+------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2018-TMI-UNet++/TestEasyDataset_eval.txt",
"chars": 479,
"preview": "+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2018-TMI-UNet++/TestHardDataset_eval.txt",
"chars": 479,
"preview": "+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2019-TPAMI-COSNet/TestEasyDataset_eval.txt",
"chars": 489,
"preview": "+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2019-TPAMI-COSNet/TestHardDataset_eval.txt",
"chars": 489,
"preview": "+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2020-AAAI-PCSA/TestEasyDataset_eval.txt",
"chars": 474,
"preview": "+-----------------+----------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | M"
},
{
"path": "eval/eval-result/2020-AAAI-PCSA/TestHardDataset_eval.txt",
"chars": 474,
"preview": "+-----------------+----------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | M"
},
{
"path": "eval/eval-result/2020-MICCAI-23DCNN/TestEasyDataset_eval.txt",
"chars": 454,
"preview": "+-----------------+--------------------+----------+-------+-----------+---------+--------+\n| Dataset | Met"
},
{
"path": "eval/eval-result/2020-MICCAI-23DCNN/TestHardDataset_eval.txt",
"chars": 454,
"preview": "+-----------------+--------------------+----------+-------+-----------+---------+--------+\n| Dataset | Met"
},
{
"path": "eval/eval-result/2020-MICCAI-ACSNet/TestEasyDataset_eval.txt",
"chars": 494,
"preview": "+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2020-MICCAI-ACSNet/TestHardDataset_eval.txt",
"chars": 494,
"preview": "+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2020-MICCAI-PraNet/TestEasyDataset_eval.txt",
"chars": 494,
"preview": "+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2020-MICCAI-PraNet/TestHardDataset_eval.txt",
"chars": 494,
"preview": "+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2020-TIP-MATNet/TestEasyDataset_eval.txt",
"chars": 479,
"preview": "+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2020-TIP-MATNet/TestHardDataset_eval.txt",
"chars": 479,
"preview": "+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2021-ICCV-DCFNet/TestEasyDataset_eval.txt",
"chars": 484,
"preview": "+-----------------+------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2021-ICCV-DCFNet/TestHardDataset_eval.txt",
"chars": 484,
"preview": "+-----------------+------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2021-ICCV-FSNet/TestEasyDataset_eval.txt",
"chars": 479,
"preview": "+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2021-ICCV-FSNet/TestHardDataset_eval.txt",
"chars": 479,
"preview": "+-----------------+-----------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2021-MICCAI-PNSNet/TestEasyDataset_eval.txt",
"chars": 494,
"preview": "+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2021-MICCAI-PNSNet/TestHardDataset_eval.txt",
"chars": 494,
"preview": "+-----------------+--------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2021-MICCAI-SANet/TestEasyDataset_eval.txt",
"chars": 489,
"preview": "+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2021-MICCAI-SANet/TestHardDataset_eval.txt",
"chars": 489,
"preview": "+-----------------+-------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2021-NIPS-AMD/TestEasyDataset_eval.txt",
"chars": 469,
"preview": "+-----------------+---------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | Me"
},
{
"path": "eval/eval-result/2021-NIPS-AMD/TestHardDataset_eval.txt",
"chars": 469,
"preview": "+-----------------+---------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | Me"
},
{
"path": "eval/eval-result/2022-TMI-PNSPlus/TestEasyDataset_eval.txt",
"chars": 484,
"preview": "+-----------------+------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval-result/2022-TMI-PNSPlus/TestHardDataset_eval.txt",
"chars": 484,
"preview": "+-----------------+------------------+----------+-------+-----------+-------+---------+--------+\n| Dataset | "
},
{
"path": "eval/eval.sh",
"chars": 1029,
"preview": "# The candidate competitors are listed here:\n# MODEL_NAMES=('2022-MIR-PNSPlus' '2021-NIPS-AMD' '2021-MICCAI-PNSNet' '202"
},
{
"path": "eval/metrics.py",
"chars": 17305,
"preview": "# -*- coding: utf-8 -*-\n# @Time : 2021/09/13\n# @Author : Johnson-Chou\n# @Email : johnson111788@gmail.com\n# @Fil"
},
{
"path": "eval/vps_evaluator.py",
"chars": 9729,
"preview": "# -*- coding: utf-8 -*-\n# @Time : 2022/03/14\n# @Author : Johnson-Chou\n# @Email : johnson111788@gmail.com\n# @Fil"
},
{
"path": "lib/__init__.py",
"chars": 0,
"preview": ""
},
{
"path": "lib/dataloader/__init__.py",
"chars": 0,
"preview": ""
},
{
"path": "lib/dataloader/dataloader.py",
"chars": 3621,
"preview": "import os\n\nimport torch\nfrom torch.utils.data import Dataset\n\nfrom scripts.config import config\nfrom lib.dataloader.prep"
},
{
"path": "lib/dataloader/preprocess.py",
"chars": 3357,
"preview": "import random\nfrom PIL import Image\n\nfrom torchvision.transforms import ToTensor as torchtotensor\n\n\nclass Compose_imglab"
},
{
"path": "lib/module/LightRFB.py",
"chars": 4322,
"preview": "import torch\nimport torch.nn as nn\n\n\nclass h_sigmoid(nn.Module):\n def __init__(self, inplace=True):\n super(h_s"
},
{
"path": "lib/module/PNS/PNS_Module/CMakeLists.txt",
"chars": 808,
"preview": "cmake_minimum_required(VERSION 3.0 FATAL_ERROR)\nproject(SA)\n\nfind_package(Torch REQUIRED)\nfind_package(CUDA REQUIRED)\n\ns"
},
{
"path": "lib/module/PNS/PNS_Module/reference.cpp",
"chars": 6504,
"preview": "#include<iostream>\n//#include<torch/torch.h>\n#include <torch/extension.h>\n#include<vector>\n\n//串行比对\nvoid sa_weight_forwar"
},
{
"path": "lib/module/PNS/PNS_Module/reference.h",
"chars": 1157,
"preview": "#ifndef REFERENCE_H_\n#define REFERENCE_H_\n\nvoid sa_weight_forward_Ref(const torch::Tensor& query,const torch::Tensor& ke"
},
{
"path": "lib/module/PNS/PNS_Module/sa.cu",
"chars": 10751,
"preview": "//#include <torch/torch.h>\n#include <torch/extension.h>\n#include <cuda_runtime.h>\n#include <iostream>\n#include <stdio.h>"
},
{
"path": "lib/module/PNS/PNS_Module/sa_ext.cpp",
"chars": 6562,
"preview": "//#include<torch/torch.h>\n#include<torch/extension.h>\n#include\"utils.h\"\n#include\"timer.h\"\n#include\"reference.h\"\n\nvoid ge"
},
{
"path": "lib/module/PNS/PNS_Module/timer.h",
"chars": 577,
"preview": "#ifndef GPU_TIMER_H__\n#define GPU_TIMER_H__\n\n#include <cuda_runtime.h>\n\nstruct GpuTimer\n{\n cudaEvent_t start;\n cudaEve"
},
{
"path": "lib/module/PNS/PNS_Module/utils.h",
"chars": 2955,
"preview": "#ifndef UTILS_H__\n#define UTILS_H__\n\n#include <iostream>\n#include <iomanip>\n#include <cuda.h>\n#include <cuda_runtime.h>\n"
},
{
"path": "lib/module/PNS/setup.py",
"chars": 718,
"preview": "from setuptools import setup\nfrom torch.utils.cpp_extension import BuildExtension, CUDAExtension\nfrom os.path import joi"
},
{
"path": "lib/module/PNSPlusModule.py",
"chars": 4364,
"preview": "import numpy as np\nfrom math import sqrt\n\nimport torch\nimport torch.nn as nn\nimport torch.nn.functional as F\nimport torc"
},
{
"path": "lib/module/PNSPlusNetwork.py",
"chars": 4810,
"preview": "import numpy as np\n\nimport torch\nimport torch.nn as nn\nimport torch.nn.functional as F\n\nfrom lib.module.LightRFB import "
},
{
"path": "lib/module/Res2Net_v1b.py",
"chars": 8183,
"preview": "import math\n\nimport torch\nimport torch.nn as nn\nimport torch.utils.model_zoo as model_zoo\n\n__all__ = ['Res2Net', 'res2ne"
},
{
"path": "lib/module/__init__.py",
"chars": 0,
"preview": ""
},
{
"path": "lib/utils/__init__.py",
"chars": 0,
"preview": ""
},
{
"path": "lib/utils/utils.py",
"chars": 628,
"preview": "def clip_gradient(optimizer, grad_clip):\n \"\"\"\n For calibrating misalignment gradient via cliping gradient techniqu"
},
{
"path": "scripts/config.py",
"chars": 1420,
"preview": "import argparse\n\n\nparser = argparse.ArgumentParser()\n\n# optimizer\nparser.add_argument('--gpu_id', type=str, default='0, "
},
{
"path": "scripts/eval_eff.py",
"chars": 1263,
"preview": "import os\nimport time\nimport numpy as np\nfrom ptflops import get_model_complexity_info\n\nimport torch\n\nfrom lib.module.PN"
},
{
"path": "scripts/my_test.py",
"chars": 4825,
"preview": "import os\nimport numpy as np\nfrom tqdm import tqdm\nfrom PIL import Image\n\nimport torch\nfrom torch.utils.data import Data"
},
{
"path": "scripts/my_train.py",
"chars": 5588,
"preview": "import os\nimport logging\nfrom datetime import datetime\n\nimport torch\nimport torch.nn as nn\nimport torch.nn.functional as"
},
{
"path": "snapshot/.placeholder",
"chars": 33,
"preview": "the snapshots will be stored here"
},
{
"path": "utils/reorganize.py",
"chars": 1745,
"preview": "import os, shutil, glob\n\nSUN_root = './data/SUN-Positive/'\nSUNSEG_root = './data/SUN-SEG-Annotation/'\n\nSUN_split_dict = "
},
{
"path": "utils/reorganize.sh",
"chars": 101,
"preview": "python ./utils/reorganize.py\n\nrm -rf ./data/SUN-Positive\nmv ./data/SUN-SEG-Annotation ./data/SUN-SEG\n"
}
]
// ... and 1 more files (download for full content)
About this extraction
This page contains the full source code of the GewelsJI/VPS GitHub repository, extracted and formatted as plain text for AI agents and large language models (LLMs). The extraction includes 68 files (263.4 KB), approximately 87.2k tokens, and a symbol index with 152 extracted functions, classes, methods, constants, and types. Use this with OpenClaw, Claude, ChatGPT, Cursor, Windsurf, or any other AI tool that accepts text input. You can copy the full output to your clipboard or download it as a .txt file.
Extracted by GitExtract — free GitHub repo to text converter for AI. Built by Nikandr Surkov.