Repository: SherryJYC/paper-MTMC Branch: main Commit: b129695d985f Files: 2 Total size: 16.1 KB Directory structure: gitextract_ofei0upc/ ├── .gitignore └── README.md ================================================ FILE CONTENTS ================================================ ================================================ FILE: .gitignore ================================================ # Byte-compiled / optimized / DLL files __pycache__/ *.py[cod] *$py.class # C extensions *.so # Distribution / packaging .Python build/ develop-eggs/ dist/ downloads/ eggs/ .eggs/ lib/ lib64/ parts/ sdist/ var/ wheels/ pip-wheel-metadata/ share/python-wheels/ *.egg-info/ .installed.cfg *.egg MANIFEST # PyInstaller # Usually these files are written by a python script from a template # before PyInstaller builds the exe, so as to inject date/other infos into it. *.manifest *.spec # Installer logs pip-log.txt pip-delete-this-directory.txt # Unit test / coverage reports htmlcov/ .tox/ .nox/ .coverage .coverage.* .cache nosetests.xml coverage.xml *.cover *.py,cover .hypothesis/ .pytest_cache/ # Translations *.mo *.pot # Django stuff: *.log local_settings.py db.sqlite3 db.sqlite3-journal # Flask stuff: instance/ .webassets-cache # Scrapy stuff: .scrapy # Sphinx documentation docs/_build/ # PyBuilder target/ # Jupyter Notebook .ipynb_checkpoints # IPython profile_default/ ipython_config.py # pyenv .python-version # pipenv # According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control. # However, in case of collaboration, if having platform-specific dependencies or dependencies # having no cross-platform support, pipenv may install dependencies that don't work, or not # install all needed dependencies. #Pipfile.lock # PEP 582; used by e.g. github.com/David-OConnor/pyflow __pypackages__/ # Celery stuff celerybeat-schedule celerybeat.pid # SageMath parsed files *.sage.py # Environments .env .venv env/ venv/ ENV/ env.bak/ venv.bak/ # Spyder project settings .spyderproject .spyproject # Rope project settings .ropeproject # mkdocs documentation /site # mypy .mypy_cache/ .dmypy.json dmypy.json # Pyre type checker .pyre/ ================================================ FILE: README.md ================================================ # MTMC A paper list of Multi Target Multi Camera (MTMC) tracking and related topics
including application case in: vehicle tracking :red_car: , pedestrian tracking :frowning_person: , sports player tracking :soccer: .
Click to show menu

1. Multi Target Single Camera Tracking Paper
2. Multi Target Multi Camera Tracking Paper
3. Related Github Repo
4. Related Competition

## Multi Target Single Camera Tracking Paper ### 2022 - Observation-Centric SORT: Rethinking SORT for Robust Multi-Object Tracking, Cao et al. [[paper]](https://arxiv.org/abs/2203.14360) [[code]](https://github.com/noahcao/OC_SORT) > interesting to see a variant of SORT (observation-centered) achieve decent results - PoserNet: Refining Relative Camera Poses Exploiting Object Detections, Taiana et al. :rainbow: [[paper]](https://arxiv.org/pdf/2207.09445.pdf) [[code]](https://github.com/IIT-PAVIS/PoserNet) > not tracking but seems applicable in MC-tracking, detect bbox from images and match roughly, use interesting GNN formulation to refine camera pose: image as node, edge as relative pose, bbox info added during message passing ### 2021 - ByteTrack: Multi-Object Tracking by Associating Every Detection Box, Zhang et al. [[paper]](https://arxiv.org/abs/2110.06864) [[code]](https://github.com/ifzhang/ByteTrack) > at first associate box with high detection score, then associate box with low detection score, improve tracking on occluded objects - Quasi-Dense Similarity Learning for Multiple Object Tracking, Pang et al. :rainbow: [[paper]](https://arxiv.org/abs/2006.06664) [[code]](https://github.com/SysCV/qdtrack) > instance similarity learning based on region proposal, flexible, no external data required - TrackFormer: Multi-Object Tracking with Transformers, Meinhardt et al. [[paper]](https://arxiv.org/abs/2101.02702) > Transformer, detection and tracking simultaneously ### 2020 - How To Train Your Deep Multi-Object Tracker, Xu et al. :rainbow: [[paper]](https://arxiv.org/abs/1906.06618) > Deep Hungarian Net, approximate MOTA, MOTP for loss function directly - Learning a Neural Solver for Multiple Object Tracking, Braso & Leal-Taixe :rainbow: [[paper]](https://arxiv.org/abs/1912.07515) > apperance embedding (node) and geometry distance embedding (edge) for graph, edge classification with cross entropy loss - Deep learning in video multi-object tracking: A survey, Ciaparrone et al. [[paper]](https://arxiv.org/abs/1907.12740) > pipeline: detection, feature extraction, affinity, association - Chained-Tracker: Chaining Paired Attentive Regression Results for End-to-End Joint Multiple-Object Detection and Tracking, Peng et al. :rainbow: [[paper]](https://arxiv.org/abs/2007.14557) [[code]](https://github.com/pjl1995/CTracker) > end-to-end MOT, use adjacent frames (chained) to combine detection, feature extraction and tracking ### 2019 - Spatial-Temporal Relation Networks for Multi-Object Tracking, Xu et al. [[paper]](https://openaccess.thecvf.com/content_ICCV_2019/papers/Xu_Spatial-Temporal_Relation_Networks_for_Multi-Object_Tracking_ICCV_2019_paper.pdf) > use appearance, location and topology cues for similarity score, then graph solved by Hungarian algorithm - Graph convolutional tracking, Gao et al. [[paper]](https://openaccess.thecvf.com/content_CVPR_2019/papers/Gao_Graph_Convolutional_Tracking_CVPR_2019_paper.pdf) > GNN, Siamese network - Tracking without bells and whistles, Bergmann et al. [[paper]](https://arxiv.org/abs/1903.05625) [[code]](https://github.com/phil-bergmann/tracking_wo_bnw) > motion and appearance extention -> Tracktor++ - Deep Learning for Visual Tracking: A Comprehensive Survey, Marvasti-Zadeh et al. [[paper]](https://arxiv.org/abs/1912.00535) > traditional and deep visual trackers - A Review of Visual Trackers and Analysis of its Application to Mobile Robot, You et al. [[paper]](https://arxiv.org/abs/1910.09761) > correlation filter, deep learning and convolutional features ### 2018 - Exploit the Connectivity: Multi-Object Tracking with TrackletNet, Wang et al. [[paper]](https://arxiv.org/abs/1811.07258) > use epipolar geometry, tracklet as node in graph - Real-time Multiple People Tracking with Deeply Learned Candidate Selection and Person Re-Identification, Chen et al. [[paper]](https://arxiv.org/abs/1809.04427)[[code]](https://github.com/longcw/MOTDT) > online MOT tracker ### 2017 - Multi-Object Tracking with Quadruplet Convolutional Neural Networks, Son et al. [[paper]](https://openaccess.thecvf.com/content_cvpr_2017/papers/Son_Multi-Object_Tracking_With_CVPR_2017_paper.pdf) > learn statistics to normalize effect of camera poses, temporal adjacent constraint for data association - Real-Time Multiple Object Tracking, Murray. [[paper]](https://www.diva-portal.org/smash/get/diva2:1146388/FULLTEXT01.pdf) > not use appearance feature, very fast, not accurate - High-Speed Tracking-by-Detection Without Using Image Information, Bochinski et al. [[paper]](http://elvera.nue.tu-berlin.de/files/1517Bochinski2017.pdf) [[code]](https://github.com/bochinski/iou-tracker) > IoU tracker, no visual cues used, fast - Online Multi-Target Tracking Using Recurrent Neural Networks, Milan et al. [[paper]](https://arxiv.org/abs/1604.03635) > RNN as tracker, LSTM for data association ### 2016 - Learning by tracking: Siamese CNN for robust target association, Leal-Taixe et al. [[paper]](https://arxiv.org/abs/1604.07866) > use Siamese CNN to learn similarity, for data association, graph solved by Linear Programming ### 2014 - Learning an image-based motion context for multiple people tracking, Leal-Taixe et al. [[paper]](https://ieeexplore.ieee.org/document/6909848) > interaction between objects, relax the dependency of tracking on detections ## Multi Target Multi Camera Tracking Paper ### 2022 - Graph Convolutional Network for Multi-Target Multi-Camera Vehicle Tracking, Luna et al. [[paper]](https://arxiv.org/pdf/2211.15538.pdf) > step 1: single camera tracking & generate appearance feature, step 2: multi camera association with GNN (single camera trajectories as node, averaged feature as node feature, cos(feature) as edge feature), weighted loss for imbalance ### 2021 - DyGLIP: A Dynamic Graph Model with Link Prediction for Accurate Multi-Camera Multiple Object Tracking, Quach et al. [[paper]](https://openaccess.thecvf.com/content/CVPR2021/papers/Quach_DyGLIP_A_Dynamic_Graph_Model_With_Link_Prediction_for_Accurate_CVPR_2021_paper.pdf) > tracklet as node, link prediction for data association, ok for w/wo overalaping view, use large training data - Online Clustering-based Multi-Camera Vehicle Tracking in Scenarios with overlapping FOVs, Luna et al. [[paper]](https://arxiv.org/pdf/2102.04091.pdf) > detection-> feature extraction, homography -> cross-camera cluster -> incremental temporal association, small latency, not very accurate ### 2020 - Real-time 3D Deep Multi-Camera Tracking, You & Jiang [[paper]](https://arxiv.org/abs/2003.11753) > fusion all views into ground-plane occupancy heatmap - City-Scale Multi-Camera Vehicle Tracking by Semantic Attribute Parsing and Cross-Camera Tracklet Matching, He et al. [[paper]](https://openaccess.thecvf.com/content_CVPRW_2020/papers/w35/He_City-Scale_Multi-Camera_Vehicle_Tracking_by_Semantic_Attribute_Parsing_and_Cross-Camera_CVPRW_2020_paper.pdf) > tracklet representation with spatial-temporal attention, then tracklet-to-target assignment - Multi-Target Multi-Camera Tracking by Tracklet-to-Target Assignment, He et al. [[paper]](https://ieeexplore.ieee.org/document/9042858) [[code]](https://github.com/GehenHe/TRACTA) > tracklet-to-target assignment - AI City Challenge 2020 – Computer Vision for Smart Transportation Applications, Chang et al. [[paper]](https://openaccess.thecvf.com/content_CVPRW_2020/papers/w35/Chang_AI_City_Challenge_2020_-_Computer_Vision_for_Smart_Transportation_CVPRW_2020_paper.pdf) > single camera tracklet -> multi-camera tracklet fusion with appearance and physical features - Multi-Camera Tracking of Vehicles based on Deep Features Re-ID and Trajectory-Based Camera Link Models, Hsu et al. [[paper]](https://openaccess.thecvf.com/content_CVPRW_2019/papers/AI%20City/Hsu_Multi-Camera_Tracking_of_Vehicles_based_on_Deep_Features_Re-ID_and_CVPRW_2019_paper.pdf) > use TrackletNet for single camera trajectory -> inter-camera tracking - ELECTRICITY: An Efficient Multi-camera Vehicle Tracking System for Intelligent City, Qian et al. [[paper]](https://openaccess.thecvf.com/content_CVPRW_2020/papers/w35/Qian_ELECTRICITY_An_Efficient_Multi-Camera_Vehicle_Tracking_System_for_Intelligent_City_CVPRW_2020_paper.pdf) > single camera tracking -> match tracklets across camera views - Pose-Assisted Multi-Camera Collaboration for Active Object Tracking, Li et al. [[paper]](https://arxiv.org/abs/2001.05161) [[code]](https://github.com/LilJing/pose-assisted-collaboration) > Reinforcement learning, collaborative multi-camera - Reconstruction of 3D flight trajectories from ad-hoc camera networks, Li et al. [[paper]](https://arxiv.org/abs/2003.04784) [[code]](https://github.com/CenekAlbl/mvus) > camera synchronization, SfM, Bundle Adjustment, spline representation for drone trajectory - The MTA Dataset for Multi Target Multi Camera Pedestrian Tracking by Weighted Distance Aggregation [[paper]](https://openaccess.thecvf.com/content_CVPRW_2020/papers/w70/Kohl_The_MTA_Dataset_for_Multi-Target_Multi-Camera_Pedestrian_Tracking_by_Weighted_CVPRW_2020_paper.pdf) > combine appearance and homography for hierachical clustering, known camera pose - Cross-View Tracking for Multi-Human 3D Pose Estimation at over 100 FPS, Chen et al. [[paper]](https://openaccess.thecvf.com/content_CVPR_2020/papers/Chen_Cross-View_Tracking_for_Multi-Human_3D_Pose_Estimation_at_Over_100_CVPR_2020_paper.pdf) ### 2019 - People tracking in multi-camera systems: a review, Iguernaissi et al. [[paper]](https://link.springer.com/article/10.1007/s11042-018-6638-5) > Centralized (combine cross-camera views before tracking, like Wen et al.) and Distributed methods (single-camera tracking before fusion) - CityFlow: A City-Scale Benchmark for Multi-Target Multi-Camera Vehicle Tracking and Re-Identification, Tang et al. [[paper]](https://arxiv.org/abs/1903.09254) - Real-Time Multi-Target Multi-Camera Tracking with Spatial-Temporal Information, Zhang & Izquierdo :rainbow: [[paper]](https://ieeexplore.ieee.org/document/8965845) > single camera detection -> create/match to track, with apperance, motion, spatial-temporal cues (cross-camera) ### 2018 - Features for Multi-Target Multi-Camera Tracking and Re-Identification, Ristani & Tomasi [[paper]](https://arxiv.org/abs/1803.10859) [[code]](https://github.com/SamvitJ/Duke-DeepCC) > tracklet -> single camera trajectory (correlation clustering) -> multi camera trajectory - Vehicle Re-Identification with the Space-Time Prior, Wu et al. [[paper]](https://openaccess.thecvf.com/content_cvpr_2018_workshops/papers/w3/Wu_Vehicle_Re-Identification_With_CVPR_2018_paper.pdf) [[code]](https://github.com/cw1204772/AIC2018_iamai) > single camera tracking -> CNN feature extraction -> multi camera tracking (KMeans) ### 2017 - Multi-Camera Multi-Target Tracking with Space-Time-View Hyper-graph, Wen et al. :rainbow: [[paper]](https://link.springer.com/article/10.1007/s11263-016-0943-0) > 3D position for affinity computation, need know camera parameters, cross-view coupling before trajectory ### 2014 - Persistent Tracking for Wide Area Aerial Surveillance, Prokaj & Medioni :rainbow: [[paper]](https://ieeexplore.ieee.org/document/6909551) > two tracker (detection and regression) in parallel, measure their correspondence ### 2013 - Hypergraphs for joint multi-view reconstruction and multi-object tracking, Hofmann et al. :rainbow: [[paper]](https://ieeexplore.ieee.org/document/6619312) [[code]](https://github.com/neohanju/HYPERGRAPH_TRACKING) > detection as node in hypergraph to find 3d reconstruction, which is node in a min-cost flow graph, solved by binary linear programming ### 2012 - Branch-and-price global optimization for multi-view multi-target tracking, Leal-Taixé et al. [[paper]](https://www.researchgate.net/publication/261200087_Branch-and-price_global_optimization_for_multi-view_multi-target_tracking) ## Related Github Repo - [Multi-camera live object tracking](https://github.com/LeonLok/Multi-Camera-Live-Object-Tracking) - [Resource collection about multi camera network](https://github.com/YanLu-nyu/Awesome-Multi-Camera-Network)
- [Recource collection about multi object tracking](https://github.com/nightmaredimple/Multi-object-Tracking-paper-code-list) - [Multi Object Tracking Paper List](https://github.com/SpyderXu/multi-object-tracking-paper-list) - [UAV detection and tracking](https://github.com/tau-adl/Detection_Tracking_JetsonTX2)
- [Resource collection about person reid dataset](https://github.com/NEU-Gou/awesome-reid-dataset)
- [OpenMMLab: toolbox for SOT, MOT](https://github.com/open-mmlab/mmtracking) - [DeepOcculusion](https://github.com/pierrebaque/DeepOcclusion) - [MOT Metrics library (Python)](https://github.com/cheind/py-motmetrics) - [MOT Metrics library (Python) 2](https://github.com/Videmo/pymot) - [Multi camera person tracker for synthetic data](https://github.com/koehlp/wda_tracker) ## Related Dataset - [Multi Track Auto (GTA)](https://github.com/schuar-iosb/mta-dataset) [[baseline provided](https://github.com/koehlp/wda_tracker)] - [BDD100K large driving dataset](https://github.com/bdd100k/bdd100k) - [Visual Tracker Benchmark](http://cvlab.hanyang.ac.kr/tracker_benchmark/datasets.html) - [DJI Drone Images](https://github.com/chuanenlin/drone-net) ## Related Competition - [AI City Challenge](https://www.aicitychallenge.org/) - [Anti-UAV Challenge](https://anti-uav.github.io/) - [Waymo Open Dataset Challenge](https://waymo.com/open/challenges) - [SoccerNet](https://www.soccer-net.org/home)