Showing preview only (218K chars total). Download the full file or copy to clipboard to get everything.
Repository: juntang-zhuang/LadderNet
Branch: master
Commit: cc440fdf799d
Files: 75
Total size: 16.3 MB
Directory structure:
gitextract_5b2xktow/
├── .gitignore
├── .idea/
│ ├── libraries/
│ │ └── R_User_Library.xml
│ ├── misc.xml
│ ├── modules.xml
│ ├── retinaunet-master.iml
│ └── workspace.xml
├── DRIVE/
│ ├── test/
│ │ └── images/
│ │ ├── 01_test.tif
│ │ ├── 02_test.tif
│ │ ├── 03_test.tif
│ │ ├── 04_test.tif
│ │ ├── 05_test.tif
│ │ ├── 06_test.tif
│ │ ├── 07_test.tif
│ │ ├── 08_test.tif
│ │ ├── 09_test.tif
│ │ ├── 10_test.tif
│ │ ├── 11_test.tif
│ │ ├── 12_test.tif
│ │ ├── 13_test.tif
│ │ ├── 14_test.tif
│ │ ├── 15_test.tif
│ │ ├── 16_test.tif
│ │ ├── 17_test.tif
│ │ ├── 18_test.tif
│ │ ├── 19_test.tif
│ │ └── 20_test.tif
│ └── training/
│ └── images/
│ ├── 21_training.tif
│ ├── 22_training.tif
│ ├── 23_training.tif
│ ├── 24_training.tif
│ ├── 25_training.tif
│ ├── 26_training.tif
│ ├── 27_training.tif
│ ├── 28_training.tif
│ ├── 29_training.tif
│ ├── 30_training.tif
│ ├── 31_training.tif
│ ├── 32_training.tif
│ ├── 33_training.tif
│ ├── 34_training.tif
│ ├── 35_training.tif
│ ├── 36_training.tif
│ ├── 37_training.tif
│ ├── 38_training.tif
│ ├── 39_training.tif
│ └── 40_training.tif
├── README.md
├── __init__.py
├── _config.yml
├── configuration.txt
├── figures/
│ └── readme
├── googlefffa84c6c8a268c9.html
├── lib/
│ ├── __init__.py
│ ├── densenet.py
│ ├── extract_patches.py
│ ├── extract_patches.py.bak
│ ├── help_functions.py
│ ├── help_functions.py.bak
│ ├── pre_processing.py
│ └── pre_processing.py.bak
├── prepare_datasets_DRIVE.py
├── src/
│ ├── LadderNetv65.py
│ ├── __init__.py
│ ├── checkpoint/
│ │ └── LadderNetv65_layer_4_filter_10.pt7
│ ├── losses.py
│ ├── retinaNN_predict.py
│ ├── retinaNN_training.py
│ └── subpixel_upscaling.py
└── test/
├── drive_performances.txt
├── performances.txt
├── test_architecture.json
├── test_best_weights.h5
├── test_configuration.txt
├── test_last_weights.h5
└── test_training.nohup
================================================
FILE CONTENTS
================================================
================================================
FILE: .gitignore
================================================
*~
*.pyc
.DS_Store
DRIVE_datasets_training_testing
================================================
FILE: .idea/libraries/R_User_Library.xml
================================================
<component name="libraryTable">
<library name="R User Library">
<CLASSES />
<SOURCES />
</library>
</component>
================================================
FILE: .idea/misc.xml
================================================
<?xml version="1.0" encoding="UTF-8"?>
<project version="4">
<component name="ProjectRootManager" version="2" project-jdk-name="Python 3.6 (py36_torch)" project-jdk-type="Python SDK" />
</project>
================================================
FILE: .idea/modules.xml
================================================
<?xml version="1.0" encoding="UTF-8"?>
<project version="4">
<component name="ProjectModuleManager">
<modules>
<module fileurl="file://$PROJECT_DIR$/.idea/retinaunet-master.iml" filepath="$PROJECT_DIR$/.idea/retinaunet-master.iml" />
</modules>
</component>
</project>
================================================
FILE: .idea/retinaunet-master.iml
================================================
<?xml version="1.0" encoding="UTF-8"?>
<module type="PYTHON_MODULE" version="4">
<component name="NewModuleRootManager">
<content url="file://$MODULE_DIR$">
<sourceFolder url="file://$MODULE_DIR$" isTestSource="false" />
<sourceFolder url="file://$MODULE_DIR$/src" isTestSource="false" />
</content>
<orderEntry type="jdk" jdkName="Python 3.6 (py36_torch)" jdkType="Python SDK" />
<orderEntry type="sourceFolder" forTests="false" />
<orderEntry type="library" name="R User Library" level="project" />
<orderEntry type="library" name="R Skeletons" level="application" />
</component>
<component name="TestRunnerService">
<option name="PROJECT_TEST_RUNNER" value="Unittests" />
</component>
</module>
================================================
FILE: .idea/workspace.xml
================================================
<?xml version="1.0" encoding="UTF-8"?>
<project version="4">
<component name="BookmarkManager">
<bookmark url="file://$PROJECT_DIR$/src/LadderNet.py" line="153" mnemonic="3" />
</component>
<component name="ChangeListManager">
<list default="true" id="ba34056a-964b-4717-aad6-9081958edce6" name="Default" comment="" />
<option name="EXCLUDED_CONVERTED_TO_IGNORED" value="true" />
<option name="SHOW_DIALOG" value="false" />
<option name="HIGHLIGHT_CONFLICTS" value="true" />
<option name="HIGHLIGHT_NON_ACTIVE_CHANGELIST" value="false" />
<option name="LAST_RESOLUTION" value="IGNORE" />
</component>
<component name="FUSProjectUsageTrigger">
<session id="-314945016">
<usages-collector id="statistics.lifecycle.project">
<counts>
<entry key="project.closed" value="1" />
<entry key="project.open.time.1" value="1" />
<entry key="project.opened" value="1" />
</counts>
</usages-collector>
<usages-collector id="statistics.file.extensions.open">
<counts>
<entry key="py" value="17" />
</counts>
</usages-collector>
<usages-collector id="statistics.file.types.open">
<counts>
<entry key="Python" value="17" />
</counts>
</usages-collector>
<usages-collector id="statistics.file.extensions.edit">
<counts>
<entry key="Python Console" value="924" />
<entry key="py" value="6229" />
<entry key="txt" value="27" />
</counts>
</usages-collector>
<usages-collector id="statistics.file.types.edit">
<counts>
<entry key="PLAIN_TEXT" value="27" />
<entry key="Python" value="7153" />
</counts>
</usages-collector>
</session>
<session id="-54659070">
<usages-collector id="statistics.lifecycle.project">
<counts>
<entry key="project.closed" value="2" />
<entry key="project.open.time.1" value="1" />
<entry key="project.open.time.8" value="1" />
<entry key="project.opened" value="2" />
</counts>
</usages-collector>
<usages-collector id="statistics.file.extensions.edit">
<counts>
<entry key="py" value="2" />
</counts>
</usages-collector>
<usages-collector id="statistics.file.types.edit">
<counts>
<entry key="Python" value="2" />
</counts>
</usages-collector>
</session>
<session id="398864315">
<usages-collector id="statistics.lifecycle.project">
<counts>
<entry key="project.closed" value="2" />
<entry key="project.open.time.0" value="2" />
<entry key="project.opened" value="2" />
</counts>
</usages-collector>
<usages-collector id="statistics.file.extensions.open">
<counts>
<entry key="txt" value="1" />
</counts>
</usages-collector>
<usages-collector id="statistics.file.types.open">
<counts>
<entry key="PLAIN_TEXT" value="1" />
</counts>
</usages-collector>
<usages-collector id="statistics.file.extensions.edit">
<counts>
<entry key="py" value="56" />
</counts>
</usages-collector>
<usages-collector id="statistics.file.types.edit">
<counts>
<entry key="Python" value="56" />
</counts>
</usages-collector>
</session>
</component>
<component name="FileEditorManager">
<leaf SIDE_TABS_SIZE_LIMIT_KEY="525">
<file pinned="false" current-in-tab="false">
<entry file="file://$PROJECT_DIR$/prepare_datasets_DRIVE.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="1422">
<caret line="79" column="64" selection-start-line="79" selection-start-column="52" selection-end-line="79" selection-end-column="64" />
<folding>
<element signature="e#183#192#0" expanded="true" />
</folding>
</state>
</provider>
</entry>
</file>
<file pinned="false" current-in-tab="false">
<entry file="file://$PROJECT_DIR$/src/retinaNN_training.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="1404">
<caret line="78" selection-start-line="78" selection-end-line="78" />
<folding>
<element signature="e#228#246#0" expanded="true" />
<marker date="1535675909209" expanded="true" signature="2906:2986" ph="..." />
</folding>
</state>
</provider>
</entry>
</file>
<file pinned="false" current-in-tab="false">
<entry file="file://$PROJECT_DIR$/configuration.txt">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="360">
<caret line="20" selection-start-line="20" selection-end-line="20" />
</state>
</provider>
</entry>
</file>
<file pinned="false" current-in-tab="false">
<entry file="file://$PROJECT_DIR$/src/LadderNet.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="5922">
<caret line="329" column="18" selection-start-line="329" selection-start-column="18" selection-end-line="329" selection-end-column="18" />
<folding>
<element signature="e#0#12#0" expanded="true" />
<marker date="1535336129873" expanded="true" signature="2478:2483" ph="..." />
<marker date="1535336129873" expanded="true" signature="9381:9390" ph="..." />
<marker date="1535336129873" expanded="true" signature="10356:10361" ph="..." />
</folding>
</state>
</provider>
</entry>
</file>
<file pinned="false" current-in-tab="false">
<entry file="file://$PROJECT_DIR$/src/LadderNetv6.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="4608">
<caret line="256" column="24" selection-start-line="256" selection-start-column="24" selection-end-line="256" selection-end-column="24" />
<folding>
<element signature="e#0#12#0" expanded="true" />
</folding>
</state>
</provider>
</entry>
</file>
<file pinned="false" current-in-tab="false">
<entry file="file://$PROJECT_DIR$/src/LadderNetv62.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="3996">
<caret line="222" column="27" selection-start-line="222" selection-start-column="27" selection-end-line="222" selection-end-column="27" />
<folding>
<element signature="e#0#12#0" expanded="true" />
</folding>
</state>
</provider>
</entry>
</file>
<file pinned="false" current-in-tab="false">
<entry file="file://$PROJECT_DIR$/src/LadderNetv63.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="3852">
<caret line="214" selection-start-line="214" selection-end-line="214" />
<folding>
<element signature="e#0#12#0" expanded="true" />
</folding>
</state>
</provider>
</entry>
</file>
<file pinned="false" current-in-tab="false">
<entry file="file://$PROJECT_DIR$/src/LadderNetv65.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="2178">
<caret line="121" selection-start-line="121" selection-end-line="121" />
<folding>
<element signature="e#0#12#0" expanded="true" />
</folding>
</state>
</provider>
</entry>
</file>
<file pinned="false" current-in-tab="false">
<entry file="file://$PROJECT_DIR$/src/LadderNetv64.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="774">
<caret line="45" column="30" selection-start-line="45" selection-start-column="30" selection-end-line="45" selection-end-column="30" />
</state>
</provider>
</entry>
</file>
<file pinned="false" current-in-tab="true">
<entry file="file://$PROJECT_DIR$/src/retinaNN_predict.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="191">
<caret line="33" column="14" lean-forward="true" selection-start-line="33" selection-start-column="14" selection-end-line="33" selection-end-column="14" />
<folding>
<element signature="e#235#253#0" expanded="true" />
</folding>
</state>
</provider>
</entry>
</file>
</leaf>
</component>
<component name="FileTemplateManagerImpl">
<option name="RECENT_TEMPLATES">
<list>
<option value="Python Script" />
</list>
</option>
</component>
<component name="FindInProjectRecents">
<findStrings>
<find>plot</find>
<find>run_GPU</find>
<find>visualize</find>
<find>path_data</find>
<find>border_mode='same'</find>
<find>compile</find>
<find>average_mode</find>
<find>self.DenseBlock</find>
<find>self.num_seg_class</find>
<find>model</find>
<find>nn.module</find>
<find>device</find>
<find>as F</find>
<find>batch_size</find>
<find>check_path</find>
<find>cudnn</find>
<find>lr_schedule</find>
<find>patches_imgs_train</find>
<find>resu</find>
<find>lr</find>
<find>resume</find>
<find>total</find>
<find>best_iou</find>
<find>RESUME</find>
</findStrings>
<replaceStrings>
<replace>padding='same</replace>
<replace>padding='same'</replace>
<replace>DenseBlock</replace>
<replace>num_seg_class</replace>
<replace>nn.Module</replace>
</replaceStrings>
</component>
<component name="IdeDocumentHistory">
<option name="CHANGED_PATHS">
<list>
<option value="$PROJECT_DIR$/run_training.py" />
<option value="$PROJECT_DIR$/lib/help_functions.py" />
<option value="$PROJECT_DIR$/src/LadderNetv3.py" />
<option value="$PROJECT_DIR$/src/LadderNetv4.py" />
<option value="$PROJECT_DIR$/lib/densenet.py" />
<option value="$PROJECT_DIR$/src/densenet.py" />
<option value="$PROJECT_DIR$/src/subpixel_upscaling.py" />
<option value="$PROJECT_DIR$/src/dense_unet.py" />
<option value="$PROJECT_DIR$/src/LadderNetv7.py" />
<option value="$PROJECT_DIR$/src/losses.py" />
<option value="$PROJECT_DIR$/src/LadderNet.py" />
<option value="$PROJECT_DIR$/src/LadderNetv6.py" />
<option value="$PROJECT_DIR$/src/LadderNetv62.py" />
<option value="$PROJECT_DIR$/configuration.txt" />
<option value="$PROJECT_DIR$/src/LaaderNetv64.py" />
<option value="$PROJECT_DIR$/src/LadderNetv63.py" />
<option value="$PROJECT_DIR$/src/LadderNetv65.py" />
<option value="$PROJECT_DIR$/src/LadderNetv64.py" />
<option value="$PROJECT_DIR$/src/retinaNN_training.py" />
<option value="$PROJECT_DIR$/src/retinaNN_predict.py" />
</list>
</option>
</component>
<component name="ProjectFrameBounds" extendedState="6">
<option name="x" value="65" />
<option name="y" value="9" />
<option name="width" value="1855" />
<option name="height" value="1084" />
</component>
<component name="ProjectView">
<navigator proportions="" version="1">
<foldersAlwaysOnTop value="true" />
</navigator>
<panes>
<pane id="Scope" />
<pane id="ProjectPane">
<subPane>
<expand>
<path>
<item name="pytorch_retina_seg_laddernet_drive" type="b2602c69:ProjectViewProjectNode" />
<item name="pytorch_retina_seg_laddernet_drive" type="462c0819:PsiDirectoryNode" />
</path>
<path>
<item name="pytorch_retina_seg_laddernet_drive" type="b2602c69:ProjectViewProjectNode" />
<item name="pytorch_retina_seg_laddernet_drive" type="462c0819:PsiDirectoryNode" />
<item name="src" type="462c0819:PsiDirectoryNode" />
</path>
</expand>
<select />
</subPane>
</pane>
</panes>
</component>
<component name="PropertiesComponent">
<property name="SearchEverywhereHistoryKey" value="	FILE	file:///media/jzkay/data1/ISLES2018/benchmark/retinaunet-master/lib/densenet.py" />
<property name="last_opened_file_path" value="$PROJECT_DIR$" />
<property name="settings.editor.selected.configurable" value="com.jetbrains.python.configuration.PyActiveSdkModuleConfigurable" />
</component>
<component name="RecentsManager">
<key name="CopyFile.RECENT_KEYS">
<recent name="$PROJECT_DIR$/src" />
</key>
</component>
<component name="RunDashboard">
<option name="ruleStates">
<list>
<RuleState>
<option name="name" value="ConfigurationTypeDashboardGroupingRule" />
</RuleState>
<RuleState>
<option name="name" value="StatusDashboardGroupingRule" />
</RuleState>
</list>
</option>
</component>
<component name="RunManager" selected="Python.retinaNN_predict">
<configuration name="prepare_datasets_DRIVE" type="PythonConfigurationType" factoryName="Python" temporary="true">
<module name="retinaunet-master" />
<option name="INTERPRETER_OPTIONS" value="" />
<option name="PARENT_ENVS" value="true" />
<envs>
<env name="PYTHONUNBUFFERED" value="1" />
</envs>
<option name="SDK_HOME" value="" />
<option name="WORKING_DIRECTORY" value="$PROJECT_DIR$" />
<option name="IS_MODULE_SDK" value="true" />
<option name="ADD_CONTENT_ROOTS" value="true" />
<option name="ADD_SOURCE_ROOTS" value="true" />
<option name="SCRIPT_NAME" value="$PROJECT_DIR$/prepare_datasets_DRIVE.py" />
<option name="PARAMETERS" value="" />
<option name="SHOW_COMMAND_LINE" value="false" />
<option name="EMULATE_TERMINAL" value="false" />
<option name="MODULE_MODE" value="false" />
<option name="REDIRECT_INPUT" value="false" />
<option name="INPUT_FILE" value="" />
<method v="2" />
</configuration>
<configuration name="retinaNN_predict" type="PythonConfigurationType" factoryName="Python" temporary="true">
<module name="retinaunet-master" />
<option name="INTERPRETER_OPTIONS" value="" />
<option name="PARENT_ENVS" value="true" />
<envs>
<env name="PYTHONUNBUFFERED" value="1" />
</envs>
<option name="SDK_HOME" value="" />
<option name="WORKING_DIRECTORY" value="$PROJECT_DIR$/src" />
<option name="IS_MODULE_SDK" value="true" />
<option name="ADD_CONTENT_ROOTS" value="true" />
<option name="ADD_SOURCE_ROOTS" value="true" />
<option name="SCRIPT_NAME" value="$PROJECT_DIR$/src/retinaNN_predict.py" />
<option name="PARAMETERS" value="" />
<option name="SHOW_COMMAND_LINE" value="false" />
<option name="EMULATE_TERMINAL" value="false" />
<option name="MODULE_MODE" value="false" />
<option name="REDIRECT_INPUT" value="false" />
<option name="INPUT_FILE" value="" />
<method v="2" />
</configuration>
<configuration name="retinaNN_training" type="PythonConfigurationType" factoryName="Python" temporary="true">
<module name="retinaunet-master" />
<option name="INTERPRETER_OPTIONS" value="" />
<option name="PARENT_ENVS" value="true" />
<envs>
<env name="PYTHONUNBUFFERED" value="1" />
</envs>
<option name="SDK_HOME" value="" />
<option name="WORKING_DIRECTORY" value="$PROJECT_DIR$/src" />
<option name="IS_MODULE_SDK" value="true" />
<option name="ADD_CONTENT_ROOTS" value="true" />
<option name="ADD_SOURCE_ROOTS" value="true" />
<option name="SCRIPT_NAME" value="$PROJECT_DIR$/src/retinaNN_training.py" />
<option name="PARAMETERS" value="" />
<option name="SHOW_COMMAND_LINE" value="false" />
<option name="EMULATE_TERMINAL" value="false" />
<option name="MODULE_MODE" value="false" />
<option name="REDIRECT_INPUT" value="false" />
<option name="INPUT_FILE" value="" />
<method v="2" />
</configuration>
<configuration name="run_training" type="PythonConfigurationType" factoryName="Python" temporary="true">
<module name="retinaunet-master" />
<option name="INTERPRETER_OPTIONS" value="" />
<option name="PARENT_ENVS" value="true" />
<envs>
<env name="PYTHONUNBUFFERED" value="1" />
</envs>
<option name="SDK_HOME" value="" />
<option name="WORKING_DIRECTORY" value="$PROJECT_DIR$" />
<option name="IS_MODULE_SDK" value="true" />
<option name="ADD_CONTENT_ROOTS" value="true" />
<option name="ADD_SOURCE_ROOTS" value="true" />
<option name="SCRIPT_NAME" value="$PROJECT_DIR$/run_training.py" />
<option name="PARAMETERS" value="" />
<option name="SHOW_COMMAND_LINE" value="false" />
<option name="EMULATE_TERMINAL" value="false" />
<option name="MODULE_MODE" value="false" />
<option name="REDIRECT_INPUT" value="false" />
<option name="INPUT_FILE" value="" />
<method v="2" />
</configuration>
<list>
<item itemvalue="Python.prepare_datasets_DRIVE" />
<item itemvalue="Python.run_training" />
<item itemvalue="Python.retinaNN_training" />
<item itemvalue="Python.retinaNN_predict" />
</list>
<recent_temporary>
<list>
<item itemvalue="Python.retinaNN_predict" />
<item itemvalue="Python.retinaNN_training" />
<item itemvalue="Python.run_training" />
<item itemvalue="Python.prepare_datasets_DRIVE" />
</list>
</recent_temporary>
</component>
<component name="SvnConfiguration">
<configuration />
</component>
<component name="TaskManager">
<task active="true" id="Default" summary="Default task">
<changelist id="ba34056a-964b-4717-aad6-9081958edce6" name="Default" comment="" />
<created>1531703474486</created>
<option name="number" value="Default" />
<option name="presentableId" value="Default" />
<updated>1531703474486</updated>
</task>
<servers />
</component>
<component name="TodoView">
<todo-panel id="selected-file">
<is-autoscroll-to-source value="true" />
</todo-panel>
<todo-panel id="all">
<are-packages-shown value="true" />
<is-autoscroll-to-source value="true" />
</todo-panel>
</component>
<component name="ToolWindowManager">
<frame x="65" y="9" width="1855" height="1084" extended-state="6" />
<editor active="true" />
<layout>
<window_info content_ui="combo" id="Project" order="0" visible="true" weight="0.18390805" />
<window_info id="Structure" order="1" side_tool="true" weight="0.25" />
<window_info id="Favorites" order="2" side_tool="true" />
<window_info anchor="bottom" id="Message" order="0" />
<window_info anchor="bottom" id="Find" order="1" />
<window_info anchor="bottom" id="Run" order="2" weight="0.49625668" />
<window_info anchor="bottom" id="Debug" order="3" weight="0.46888888" />
<window_info anchor="bottom" id="Cvs" order="4" weight="0.25" />
<window_info anchor="bottom" id="Inspection" order="5" weight="0.4" />
<window_info anchor="bottom" id="TODO" order="6" weight="0.32972974" />
<window_info anchor="bottom" id="Event Log" order="7" side_tool="true" />
<window_info anchor="bottom" id="Version Control" order="8" show_stripe_button="false" />
<window_info anchor="bottom" id="Python Console" order="9" weight="0.32972974" />
<window_info anchor="bottom" id="Terminal" order="10" weight="0.32972974" />
<window_info anchor="right" id="R Graphics" order="0" />
<window_info anchor="right" id="Commander" order="1" weight="0.4" />
<window_info anchor="right" id="R Packages" order="2" />
<window_info anchor="right" id="Ant Build" order="3" weight="0.25" />
<window_info anchor="right" content_ui="combo" id="Hierarchy" order="4" weight="0.25" />
</layout>
</component>
<component name="VcsContentAnnotationSettings">
<option name="myLimit" value="2678400000" />
</component>
<component name="editorHistoryManager">
<entry file="file://$USER_HOME$/miniconda3/envs/py36/lib/python3.6/configparser.py" />
<entry file="file://$USER_HOME$/miniconda3/envs/py36/lib/python3.6/site-packages/keras/layers/core.py" />
<entry file="file://$USER_HOME$/miniconda3/envs/py36/lib/python3.6/site-packages/sklearn/metrics/ranking.py" />
<entry file="file://$PROJECT_DIR$/src/densenet.py" />
<entry file="file://$PROJECT_DIR$/src/subpixel_upscaling.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="320">
<caret line="78" column="69" selection-start-line="78" selection-start-column="69" selection-end-line="78" selection-end-column="69" />
</state>
</provider>
</entry>
<entry file="file://$USER_HOME$/miniconda3/envs/py36/lib/python3.6/site-packages/keras/layers/convolutional.py" />
<entry file="file://$PROJECT_DIR$/src/dense_unet.py" />
<entry file="file://$PROJECT_DIR$/lib/densenet.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="725">
<caret line="1017" selection-start-line="1017" selection-end-line="1017" />
</state>
</provider>
</entry>
<entry file="file://$PROJECT_DIR$/src/LadderNetv3.py" />
<entry file="file://$PROJECT_DIR$/src/LadderNetv4.py" />
<entry file="file://$PROJECT_DIR$/src/LadderNetv7.py" />
<entry file="file://$PROJECT_DIR$/run_training.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="774">
<caret line="44" column="37" selection-start-line="44" selection-start-column="37" selection-end-line="44" selection-end-column="37" />
</state>
</provider>
</entry>
<entry file="file://$PROJECT_DIR$/lib/help_functions.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="156">
<caret line="77" selection-start-line="77" selection-end-line="77" />
</state>
</provider>
</entry>
<entry file="file://$PROJECT_DIR$/lib/pre_processing.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="330">
<caret line="15" column="4" selection-start-line="15" selection-start-column="4" selection-end-line="15" selection-end-column="4" />
</state>
</provider>
</entry>
<entry file="file://$PROJECT_DIR$/lib/extract_patches.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="425">
<caret line="282" column="48" selection-start-line="282" selection-start-column="48" selection-end-line="282" selection-end-column="48" />
</state>
</provider>
</entry>
<entry file="file://$USER_HOME$/miniconda3/envs/py36_torch/lib/python3.6/site-packages/torch/nn/modules/module.py" />
<entry file="file://$USER_HOME$/miniconda3/envs/py36_torch/lib/python3.6/site-packages/torch/nn/functional.py" />
<entry file="file://$USER_HOME$/miniconda3/envs/py36_torch/lib/python3.6/site-packages/torch/nn/modules/conv.py" />
<entry file="file://$PROJECT_DIR$/src/LaaderNetv64.py" />
<entry file="file://$PROJECT_DIR$/src/losses.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="324">
<caret line="20" column="17" selection-start-line="20" selection-start-column="17" selection-end-line="20" selection-end-column="17" />
<folding>
<element signature="e#0#12#0" expanded="true" />
</folding>
</state>
</provider>
</entry>
<entry file="file://$PROJECT_DIR$/prepare_datasets_DRIVE.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="1422">
<caret line="79" column="64" selection-start-line="79" selection-start-column="52" selection-end-line="79" selection-end-column="64" />
<folding>
<element signature="e#183#192#0" expanded="true" />
</folding>
</state>
</provider>
</entry>
<entry file="file://$PROJECT_DIR$/src/retinaNN_training.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="1404">
<caret line="78" selection-start-line="78" selection-end-line="78" />
<folding>
<element signature="e#228#246#0" expanded="true" />
<marker date="1535675909209" expanded="true" signature="2906:2986" ph="..." />
</folding>
</state>
</provider>
</entry>
<entry file="file://$PROJECT_DIR$/configuration.txt">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="360">
<caret line="20" selection-start-line="20" selection-end-line="20" />
</state>
</provider>
</entry>
<entry file="file://$PROJECT_DIR$/src/LadderNet.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="5922">
<caret line="329" column="18" selection-start-line="329" selection-start-column="18" selection-end-line="329" selection-end-column="18" />
<folding>
<element signature="e#0#12#0" expanded="true" />
<marker date="1535336129873" expanded="true" signature="2478:2483" ph="..." />
<marker date="1535336129873" expanded="true" signature="9381:9390" ph="..." />
<marker date="1535336129873" expanded="true" signature="10356:10361" ph="..." />
</folding>
</state>
</provider>
</entry>
<entry file="file://$PROJECT_DIR$/src/LadderNetv6.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="4608">
<caret line="256" column="24" selection-start-line="256" selection-start-column="24" selection-end-line="256" selection-end-column="24" />
<folding>
<element signature="e#0#12#0" expanded="true" />
</folding>
</state>
</provider>
</entry>
<entry file="file://$PROJECT_DIR$/src/LadderNetv62.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="3996">
<caret line="222" column="27" selection-start-line="222" selection-start-column="27" selection-end-line="222" selection-end-column="27" />
<folding>
<element signature="e#0#12#0" expanded="true" />
</folding>
</state>
</provider>
</entry>
<entry file="file://$PROJECT_DIR$/src/LadderNetv63.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="3852">
<caret line="214" selection-start-line="214" selection-end-line="214" />
<folding>
<element signature="e#0#12#0" expanded="true" />
</folding>
</state>
</provider>
</entry>
<entry file="file://$PROJECT_DIR$/src/LadderNetv65.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="2178">
<caret line="121" selection-start-line="121" selection-end-line="121" />
<folding>
<element signature="e#0#12#0" expanded="true" />
</folding>
</state>
</provider>
</entry>
<entry file="file://$PROJECT_DIR$/src/LadderNetv64.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="774">
<caret line="45" column="30" selection-start-line="45" selection-start-column="30" selection-end-line="45" selection-end-column="30" />
</state>
</provider>
</entry>
<entry file="file://$PROJECT_DIR$/src/retinaNN_predict.py">
<provider selected="true" editor-type-id="text-editor">
<state relative-caret-position="191">
<caret line="33" column="14" lean-forward="true" selection-start-line="33" selection-start-column="14" selection-end-line="33" selection-end-column="14" />
<folding>
<element signature="e#235#253#0" expanded="true" />
</folding>
</state>
</provider>
</entry>
</component>
</project>
================================================
FILE: README.md
================================================
* Implementation for paper [LadderNet: Multi-path networks based on U-Net for medical image segmentation
](https://arxiv.org/abs/1810.07810)
* This implementation is based on [orobix implementation](https://github.com/orobix/retina-unet). Main difference is the structure of the model.
# Requirement
* Python3.6
* PyTorch 0.4
* configparser
# How to run
* run ```python prepare_datasets_DRIVE.py``` to generate hdf5 file of training data
* run ```cd src```
* run ```python retinaNN_training.py``` to train
* run ```python retinaNN_predict.py``` to test
# Parameter defination
* parameters (path, patch size, et al.) are defined in <b>"configuration.txt"</b>
* training parameters are defined in src/retinaNN_training.py line 49 t 84 with notes <b>"=====Define parameters here =========" </b>
# Pretrained weights
* pretrained weights are stored in <b>"src/checkpoint"</b>
* results are stored in <b>"test/"</b>
# Results
The results reported in the `./test` folder are referred to the trained model which reported the minimum validation loss. The `./test` folder includes:
- Model:
- `test_model.png` schematic representation of the neural network
- `test_architecture.json` description of the model in json format
- `test_best_weights.h5` weights of the model which reported the minimum validation loss, as HDF5 file
- `test_last_weights.h5` weights of the model at last epoch (150th), as HDF5 file
- `test_configuration.txt` configuration of the parameters of the experiment
- Experiment results:
- `performances.txt` summary of the test results, including the confusion matrix
- `Precision_recall.png` the precision-recall plot and the corresponding Area Under the Curve (AUC)
- `ROC.png` the Receiver Operating Characteristic (ROC) curve and the corresponding AUC
- `all_*.png` the 20 images of the pre-processed originals, ground truth and predictions relative to the DRIVE testing dataset
- `sample_input_*.png` sample of 40 patches of the pre-processed original training images and the corresponding ground truth
- `test_Original_GroundTruth_Prediction*.png` from top to bottom, the original pre-processed image, the ground truth and the prediction. In the predicted image, each pixel shows the vessel predicted probability, no threshold is applied.
The following table compares this method to other recent techniques, which have published their performance in terms of Area Under the ROC curve (AUC ROC) on the DRIVE dataset.
| Method | AUC ROC on DRIVE |
| ----------------------- |:----------------:|
| Soares et al [1] | .9614 |
| Azzopardi et al. [2] | .9614 |
| Osareh et al [3] | .9650 |
| Roychowdhury et al. [4] | .9670 |
| Fraz et al. [5] | .9747 |
| Qiaoliang et al. [6] | .9738 |
| Melinscak et al. [7] | .9749 |
| Liskowski et al.^ [8] | .9790 |
| orobix | .9790 |
| **this method** | **.9794** |
================================================
FILE: __init__.py
================================================
================================================
FILE: _config.yml
================================================
theme: jekyll-theme-cayman
================================================
FILE: configuration.txt
================================================
[data paths]
path_local = ../DRIVE_datasets_training_testing/
train_imgs_original = DRIVE_dataset_imgs_train.hdf5
train_groundTruth = DRIVE_dataset_groundTruth_train.hdf5
train_border_masks = DRIVE_dataset_borderMasks_train.hdf5
test_imgs_original = DRIVE_dataset_imgs_test.hdf5
test_groundTruth = DRIVE_dataset_groundTruth_test.hdf5
test_border_masks = DRIVE_dataset_borderMasks_test.hdf5
[experiment name]
name = test
[data attributes]
#Dimensions of the patches extracted from the full images
patch_height = 48
patch_width = 48
[training settings]
#number of total patches:
N_subimgs = 190000
#if patches are extracted only inside the field of view:
inside_FOV = False
#Number of training epochs
N_epochs = 150
batch_size = 1024
#if running with nohup
nohup = True
[testing settings]
#Choose the model to test: best==epoch with min loss, last==last epoch
best_last = best
#number of full images for the test (max 20)
full_images_to_test = 20
#How many original-groundTruth-prediction images are visualized in each image
N_group_visual = 1
#Compute average in the prediction, improve results but require more patches to be predicted
average_mode = True
#Only if average_mode==True. Stride for patch extraction, lower value require more patches to be predicted
stride_height = 5
stride_width = 5
#if running with nohup
nohup = False
================================================
FILE: figures/readme
================================================
test
================================================
FILE: googlefffa84c6c8a268c9.html
================================================
google-site-verification: googlefffa84c6c8a268c9.html
================================================
FILE: lib/__init__.py
================================================
================================================
FILE: lib/densenet.py
================================================
# -*- coding: utf-8 -*-
'''
code from Keras.contrib. This is just a local copy in case that repo changes.
DenseNet and DenseNet-FCN models for Keras.
DenseNet is a network architecture where each layer is directly connected
to every other layer in a feed-forward fashion (within each dense block).
For each layer, the feature maps of all preceding layers are treated as
separate inputs whereas its own feature maps are passed on as inputs to
all subsequent layers. This connectivity pattern yields state-of-the-art
accuracies on CIFAR10/100 (with or without data augmentation) and SVHN.
On the large scale ILSVRC 2012 (ImageNet) dataset, DenseNet achieves a
similar accuracy as ResNet, but using less than half the amount of
parameters and roughly half the number of FLOPs.
DenseNets support any input image size of 32x32 or greater, and are thus
suited for CIFAR-10 or CIFAR-100 datasets. There are two types of DenseNets,
one suited for smaller images (DenseNet) and one suited for ImageNet,
called DenseNetImageNet. They are differentiated by the strided convolution
and pooling operations prior to the initial dense block.
The following table describes the size and accuracy of DenseNetImageNet models
on the ImageNet dataset (single crop), for which weights are provided:
------------------------------------------------------------------------------------
Model type | ImageNet Acc (Top 1) | ImageNet Acc (Top 5) | Params (M) |
------------------------------------------------------------------------------------
| DenseNet-121 | 25.02 % | 7.71 % | 8.0 |
| DenseNet-169 | 23.80 % | 6.85 % | 14.3 |
| DenseNet-201 | 22.58 % | 6.34 % | 20.2 |
| DenseNet-161 | 22.20 % | - % | 28.9 |
------------------------------------------------------------------------------------
DenseNets can be extended to image segmentation tasks as described in the
paper "The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for
Semantic Segmentation". Here, the dense blocks are arranged and concatenated
with long skip connections for state of the art performance on the CamVid dataset.
# Reference
- [Densely Connected Convolutional Networks](https://arxiv.org/pdf/1608.06993.pdf)
- [The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation]
(https://arxiv.org/pdf/1611.09326.pdf)
This implementation is based on the following reference code:
- https://github.com/gpleiss/efficient_densenet_pytorch
- https://github.com/liuzhuang13/DenseNet
'''
from __future__ import print_function
from __future__ import absolute_import
from __future__ import division
import warnings
from keras.models import Model
from keras.layers import Dense
from keras.layers import Dropout
from keras.layers import Activation
from keras.layers import Reshape
from keras.layers import Conv2D
from keras.layers import Conv2DTranspose
from keras.layers import UpSampling2D
from keras.layers import MaxPooling2D
from keras.layers import AveragePooling2D
from keras.layers import GlobalMaxPooling2D
from keras.layers import GlobalAveragePooling2D
from keras.layers import Input
from keras.layers import concatenate
from keras.layers import BatchNormalization
from keras.regularizers import l2
from keras.utils.layer_utils import convert_all_kernels_in_model
from keras.utils.data_utils import get_file
from keras.engine.topology import get_source_inputs
from keras.applications.imagenet_utils import _obtain_input_shape
from keras.applications.imagenet_utils import decode_predictions
from keras.applications.imagenet_utils import preprocess_input as _preprocess_input
import keras.backend as K
from keras.callbacks import ModelCheckpoint, CSVLogger, EarlyStopping, ReduceLROnPlateau
from subpixel_upscaling import SubPixelUpscaling
DENSENET_121_WEIGHTS_PATH = r'https://github.com/titu1994/DenseNet/releases/download/v3.0/DenseNet-BC-121-32.h5'
DENSENET_161_WEIGHTS_PATH = r'https://github.com/titu1994/DenseNet/releases/download/v3.0/DenseNet-BC-161-48.h5'
DENSENET_169_WEIGHTS_PATH = r'https://github.com/titu1994/DenseNet/releases/download/v3.0/DenseNet-BC-169-32.h5'
DENSENET_121_WEIGHTS_PATH_NO_TOP = r'https://github.com/titu1994/DenseNet/releases/download/v3.0/DenseNet-BC-121-32-no-top.h5'
DENSENET_161_WEIGHTS_PATH_NO_TOP = r'https://github.com/titu1994/DenseNet/releases/download/v3.0/DenseNet-BC-161-48-no-top.h5'
DENSENET_169_WEIGHTS_PATH_NO_TOP = r'https://github.com/titu1994/DenseNet/releases/download/v3.0/DenseNet-BC-169-32-no-top.h5'
def preprocess_input(x, data_format=None):
"""Preprocesses a tensor encoding a batch of images.
# Arguments
x: input Numpy tensor, 4D.
data_format: data format of the image tensor.
# Returns
Preprocessed tensor.
"""
x = _preprocess_input(x, data_format=data_format)
x *= 0.017 # scale values
return x
def DenseNet(input_shape=None,
depth=40,
nb_dense_block=3,
growth_rate=12,
nb_filter=-1,
nb_layers_per_block=-1,
bottleneck=False,
reduction=0.0,
dropout_rate=0.0,
weight_decay=1e-4,
subsample_initial_block=False,
include_top=True,
weights=None,
input_tensor=None,
pooling=None,
classes=10,
activation='softmax',
transition_pooling='avg'):
'''Instantiate the DenseNet architecture.
The model and the weights are compatible with both
TensorFlow and Theano. The dimension ordering
convention used by the model is the one
specified in your Keras config file.
# Arguments
input_shape: optional shape tuple, only to be specified
if `include_top` is False (otherwise the input shape
has to be `(224, 224, 3)` (with `channels_last` dim ordering)
or `(3, 224, 224)` (with `channels_first` dim ordering).
It should have exactly 3 inputs channels,
and width and height should be no smaller than 8.
E.g. `(224, 224, 3)` would be one valid value.
depth: number or layers in the DenseNet
nb_dense_block: number of dense blocks to add to end
growth_rate: number of filters to add per dense block
nb_filter: initial number of filters. -1 indicates initial
number of filters will default to 2 * growth_rate
nb_layers_per_block: number of layers in each dense block.
Can be a -1, positive integer or a list.
If -1, calculates nb_layer_per_block from the network depth.
If positive integer, a set number of layers per dense block.
If list, nb_layer is used as provided. Note that list size must
be nb_dense_block
bottleneck: flag to add bottleneck blocks in between dense blocks
reduction: reduction factor of transition blocks.
Note : reduction value is inverted to compute compression.
dropout_rate: dropout rate
weight_decay: weight decay rate
subsample_initial_block: Changes model type to suit different datasets.
Should be set to True for ImageNet, and False for CIFAR datasets.
When set to True, the initial convolution will be strided and
adds a MaxPooling2D before the initial dense block.
include_top: whether to include the fully-connected
layer at the top of the network.
weights: one of `None` (random initialization) or
'imagenet' (pre-training on ImageNet)..
input_tensor: optional Keras tensor (i.e. output of `layers.Input()`)
to use as image input for the model.
pooling: Optional pooling mode for feature extraction
when `include_top` is `False`.
- `None` means that the output of the model
will be the 4D tensor output of the
last convolutional layer.
- `avg` means that global average pooling
will be applied to the output of the
last convolutional layer, and thus
the output of the model will be a
2D tensor.
- `max` means that global max pooling will
be applied.
classes: optional number of classes to classify images
into, only to be specified if `include_top` is True, and
if no `weights` argument is specified.
activation: Type of activation at the top layer. Can be one of
'softmax' or 'sigmoid'. Note that if sigmoid is used,
classes must be 1.
transition_pooling: `avg` for avg pooling (default), `max` for max pooling,
None for no pooling during scale transition blocks. Please note that this
default differs from the DenseNetFCN paper in accordance with the DenseNet
paper.
# Returns
A Keras model instance.
# Raises
ValueError: in case of invalid argument for `weights`,
or invalid input shape.
'''
if weights not in {'imagenet', None}:
raise ValueError('The `weights` argument should be either '
'`None` (random initialization) or `imagenet` '
'(pre-training on ImageNet).')
if weights == 'imagenet' and include_top and classes != 1000:
raise ValueError('If using `weights` as ImageNet with `include_top` '
'as true, `classes` should be 1000')
if activation not in ['softmax', 'sigmoid']:
raise ValueError('activation must be one of "softmax" or "sigmoid"')
if activation == 'sigmoid' and classes != 1:
raise ValueError('sigmoid activation can only be used when classes = 1')
# Determine proper input shape
input_shape = _obtain_input_shape(input_shape,
default_size=32,
min_size=8,
data_format=K.image_data_format(),
require_flatten=include_top)
if input_tensor is None:
img_input = Input(shape=input_shape)
else:
if not K.is_keras_tensor(input_tensor):
img_input = Input(tensor=input_tensor, shape=input_shape)
else:
img_input = input_tensor
x = __create_dense_net(classes, img_input, include_top, depth, nb_dense_block,
growth_rate, nb_filter, nb_layers_per_block, bottleneck,
reduction, dropout_rate, weight_decay, subsample_initial_block,
pooling, activation, transition_pooling)
# Ensure that the model takes into account
# any potential predecessors of `input_tensor`.
if input_tensor is not None:
inputs = get_source_inputs(input_tensor)
else:
inputs = img_input
# Create model.
model = Model(inputs, x, name='densenet')
# load weights
if weights == 'imagenet':
weights_loaded = False
if (depth == 121) and (nb_dense_block == 4) and (growth_rate == 32) and (nb_filter == 64) and \
(bottleneck is True) and (reduction == 0.5) and (subsample_initial_block):
if include_top:
weights_path = get_file('DenseNet-BC-121-32.h5',
DENSENET_121_WEIGHTS_PATH,
cache_subdir='models',
md5_hash='a439dd41aa672aef6daba4ee1fd54abd')
else:
weights_path = get_file('DenseNet-BC-121-32-no-top.h5',
DENSENET_121_WEIGHTS_PATH_NO_TOP,
cache_subdir='models',
md5_hash='55e62a6358af8a0af0eedf399b5aea99')
model.load_weights(weights_path, by_name=True)
weights_loaded = True
if (depth == 161) and (nb_dense_block == 4) and (growth_rate == 48) and (nb_filter == 96) and \
(bottleneck is True) and (reduction == 0.5) and (subsample_initial_block):
if include_top:
weights_path = get_file('DenseNet-BC-161-48.h5',
DENSENET_161_WEIGHTS_PATH,
cache_subdir='models',
md5_hash='6c326cf4fbdb57d31eff04333a23fcca')
else:
weights_path = get_file('DenseNet-BC-161-48-no-top.h5',
DENSENET_161_WEIGHTS_PATH_NO_TOP,
cache_subdir='models',
md5_hash='1a9476b79f6b7673acaa2769e6427b92')
model.load_weights(weights_path, by_name=True)
weights_loaded = True
if (depth == 169) and (nb_dense_block == 4) and (growth_rate == 32) and (nb_filter == 64) and \
(bottleneck is True) and (reduction == 0.5) and (subsample_initial_block):
if include_top:
weights_path = get_file('DenseNet-BC-169-32.h5',
DENSENET_169_WEIGHTS_PATH,
cache_subdir='models',
md5_hash='914869c361303d2e39dec640b4e606a6')
else:
weights_path = get_file('DenseNet-BC-169-32-no-top.h5',
DENSENET_169_WEIGHTS_PATH_NO_TOP,
cache_subdir='models',
md5_hash='89c19e8276cfd10585d5fadc1df6859e')
model.load_weights(weights_path, by_name=True)
weights_loaded = True
if weights_loaded:
if K.backend() == 'theano':
convert_all_kernels_in_model(model)
if K.image_data_format() == 'channels_first' and K.backend() == 'tensorflow':
warnings.warn('You are using the TensorFlow backend, yet you '
'are using the Theano '
'image data format convention '
'(`image_data_format="channels_first"`). '
'For best performance, set '
'`image_data_format="channels_last"` in '
'your Keras config '
'at ~/.keras/keras.json.')
print("Weights for the model were loaded successfully")
return model
def DenseNetFCN(input_shape, nb_dense_block=5, growth_rate=16, nb_layers_per_block=4,
reduction=0.0, dropout_rate=0.2, weight_decay=1E-4, init_conv_filters=48,
include_top=True, weights=None, input_tensor=None, classes=1, activation='sigmoid',
upsampling_conv=128, upsampling_type='deconv', early_transition=False,
transition_pooling='max', initial_kernel_size=(3, 3)):
'''Instantiate the DenseNet FCN architecture.
Note that when using TensorFlow,
for best performance you should set
`image_data_format='channels_last'` in your Keras config
at ~/.keras/keras.json.
# Arguments
nb_dense_block: number of dense blocks to add to end (generally = 3)
growth_rate: number of filters to add per dense block
nb_layers_per_block: number of layers in each dense block.
Can be a positive integer or a list.
If positive integer, a set number of layers per dense block.
If list, nb_layer is used as provided. Note that list size must
be (nb_dense_block + 1)
reduction: reduction factor of transition blocks.
Note : reduction value is inverted to compute compression.
dropout_rate: dropout rate
weight_decay: weight decay factor
init_conv_filters: number of layers in the initial convolution layer
include_top: whether to include the fully-connected
layer at the top of the network.
weights: one of `None` (random initialization) or
'cifar10' (pre-training on CIFAR-10)..
input_tensor: optional Keras tensor (i.e. output of `layers.Input()`)
to use as image input for the model.
input_shape: optional shape tuple, only to be specified
if `include_top` is False (otherwise the input shape
has to be `(32, 32, 3)` (with `channels_last` dim ordering)
or `(3, 32, 32)` (with `channels_first` dim ordering).
It should have exactly 3 inputs channels,
and width and height should be no smaller than 8.
E.g. `(200, 200, 3)` would be one valid value.
classes: optional number of classes to classify images
into, only to be specified if `include_top` is True, and
if no `weights` argument is specified.
activation: Type of activation at the top layer. Can be one of 'softmax' or 'sigmoid'.
Note that if sigmoid is used, classes must be 1.
upsampling_conv: number of convolutional layers in upsampling via subpixel convolution
upsampling_type: Can be one of 'deconv', 'upsampling' and
'subpixel'. Defines type of upsampling algorithm used.
batchsize: Fixed batch size. This is a temporary requirement for
computation of output shape in the case of Deconvolution2D layers.
Parameter will be removed in next iteration of Keras, which infers
output shape of deconvolution layers automatically.
early_transition: Start with an extra initial transition down and end with an extra
transition up to reduce the network size.
initial_kernel_size: The first Conv2D kernel might vary in size based on the
application, this parameter makes it configurable.
# Returns
A Keras model instance.
'''
if weights not in {None}:
raise ValueError('The `weights` argument should be '
'`None` (random initialization) as no '
'model weights are provided.')
upsampling_type = upsampling_type.lower()
if upsampling_type not in ['upsampling', 'deconv', 'subpixel']:
raise ValueError('Parameter "upsampling_type" must be one of "upsampling", '
'"deconv" or "subpixel".')
if input_shape is None:
raise ValueError('For fully convolutional models, input shape must be supplied.')
if type(nb_layers_per_block) is not list and nb_dense_block < 1:
raise ValueError('Number of dense layers per block must be greater than 1. Argument '
'value was %d.' % (nb_layers_per_block))
if activation not in ['softmax', 'sigmoid']:
raise ValueError('activation must be one of "softmax" or "sigmoid"')
if activation == 'sigmoid' and classes != 1:
raise ValueError('sigmoid activation can only be used when classes = 1')
# Determine proper input shape
min_size = 2 ** nb_dense_block
if K.image_data_format() == 'channels_first':
if input_shape is not None:
if ((input_shape[1] is not None and input_shape[1] < min_size) or
(input_shape[2] is not None and input_shape[2] < min_size)):
raise ValueError('Input size must be at least ' +
str(min_size) + 'x' + str(min_size) + ', got '
'`input_shape=' + str(input_shape) + '`')
else:
input_shape = (classes, None, None)
else:
if input_shape is not None:
if ((input_shape[0] is not None and input_shape[0] < min_size) or
(input_shape[1] is not None and input_shape[1] < min_size)):
raise ValueError('Input size must be at least ' +
str(min_size) + 'x' + str(min_size) + ', got '
'`input_shape=' + str(input_shape) + '`')
else:
input_shape = (None, None, classes)
if input_tensor is None:
img_input = Input(shape=input_shape)
else:
if not K.is_keras_tensor(input_tensor):
img_input = Input(tensor=input_tensor, shape=input_shape)
else:
img_input = input_tensor
x = __create_fcn_dense_net(classes, img_input, include_top, nb_dense_block, growth_rate,
reduction, dropout_rate, weight_decay,
nb_layers_per_block, upsampling_conv, upsampling_type,
init_conv_filters, input_shape, activation,
early_transition, transition_pooling, initial_kernel_size)
# Ensure that the model takes into account
# any potential predecessors of `input_tensor`.
if input_tensor is not None:
inputs = get_source_inputs(input_tensor)
else:
inputs = img_input
# Create model.
model = Model(inputs, x, name='fcn-densenet')
return model
def DenseNetImageNet121(input_shape=None,
bottleneck=True,
reduction=0.5,
dropout_rate=0.0,
weight_decay=1e-4,
include_top=True,
weights='imagenet',
input_tensor=None,
pooling=None,
classes=1000,
activation='softmax'):
return DenseNet(input_shape, depth=121, nb_dense_block=4, growth_rate=32, nb_filter=64,
nb_layers_per_block=[6, 12, 24, 16], bottleneck=bottleneck, reduction=reduction,
dropout_rate=dropout_rate, weight_decay=weight_decay, subsample_initial_block=True,
include_top=include_top, weights=weights, input_tensor=input_tensor,
pooling=pooling, classes=classes, activation=activation)
def DenseNetImageNet169(input_shape=None,
bottleneck=True,
reduction=0.5,
dropout_rate=0.0,
weight_decay=1e-4,
include_top=True,
weights='imagenet',
input_tensor=None,
pooling=None,
classes=1000,
activation='softmax'):
return DenseNet(input_shape, depth=169, nb_dense_block=4, growth_rate=32, nb_filter=64,
nb_layers_per_block=[6, 12, 32, 32], bottleneck=bottleneck, reduction=reduction,
dropout_rate=dropout_rate, weight_decay=weight_decay, subsample_initial_block=True,
include_top=include_top, weights=weights, input_tensor=input_tensor,
pooling=pooling, classes=classes, activation=activation)
def DenseNetImageNet201(input_shape=None,
bottleneck=True,
reduction=0.5,
dropout_rate=0.0,
weight_decay=1e-4,
include_top=True,
weights=None,
input_tensor=None,
pooling=None,
classes=1000,
activation='softmax'):
return DenseNet(input_shape, depth=201, nb_dense_block=4, growth_rate=32, nb_filter=64,
nb_layers_per_block=[6, 12, 48, 32], bottleneck=bottleneck, reduction=reduction,
dropout_rate=dropout_rate, weight_decay=weight_decay, subsample_initial_block=True,
include_top=include_top, weights=weights, input_tensor=input_tensor,
pooling=pooling, classes=classes, activation=activation)
def DenseNetImageNet264(input_shape=None,
bottleneck=True,
reduction=0.5,
dropout_rate=0.0,
weight_decay=1e-4,
include_top=True,
weights=None,
input_tensor=None,
pooling=None,
classes=1000,
activation='softmax'):
return DenseNet(input_shape, depth=201, nb_dense_block=4, growth_rate=32, nb_filter=64,
nb_layers_per_block=[6, 12, 64, 48], bottleneck=bottleneck, reduction=reduction,
dropout_rate=dropout_rate, weight_decay=weight_decay, subsample_initial_block=True,
include_top=include_top, weights=weights, input_tensor=input_tensor,
pooling=pooling, classes=classes, activation=activation)
def DenseNetImageNet161(input_shape=None,
bottleneck=True,
reduction=0.5,
dropout_rate=0.0,
weight_decay=1e-4,
include_top=True,
weights='imagenet',
input_tensor=None,
pooling=None,
classes=1000,
activation='softmax'):
return DenseNet(input_shape, depth=161, nb_dense_block=4, growth_rate=48, nb_filter=96,
nb_layers_per_block=[6, 12, 36, 24], bottleneck=bottleneck, reduction=reduction,
dropout_rate=dropout_rate, weight_decay=weight_decay, subsample_initial_block=True,
include_top=include_top, weights=weights, input_tensor=input_tensor,
pooling=pooling, classes=classes, activation=activation)
def name_or_none(prefix, name):
return prefix + name if (prefix is not None and name is not None) else None
def __conv_block(ip, nb_filter, bottleneck=False, dropout_rate=None, weight_decay=1e-4, block_prefix=None):
'''
Adds a convolution layer (with batch normalization and relu),
and optionally a bottleneck layer.
# Arguments
ip: Input tensor
nb_filter: integer, the dimensionality of the output space
(i.e. the number output of filters in the convolution)
bottleneck: if True, adds a bottleneck convolution block
dropout_rate: dropout rate
weight_decay: weight decay factor
block_prefix: str, for unique layer naming
# Input shape
4D tensor with shape:
`(samples, channels, rows, cols)` if data_format='channels_first'
or 4D tensor with shape:
`(samples, rows, cols, channels)` if data_format='channels_last'.
# Output shape
4D tensor with shape:
`(samples, filters, new_rows, new_cols)` if data_format='channels_first'
or 4D tensor with shape:
`(samples, new_rows, new_cols, filters)` if data_format='channels_last'.
`rows` and `cols` values might have changed due to stride.
# Returns
output tensor of block
'''
with K.name_scope('ConvBlock'):
concat_axis = 1 if K.image_data_format() == 'channels_first' else -1
x = BatchNormalization(axis=concat_axis, epsilon=1.1e-5, name=name_or_none(block_prefix, '_bn'))(ip)
x = Activation('relu')(x)
if bottleneck:
inter_channel = nb_filter * 4
x = Conv2D(inter_channel, (1, 1), kernel_initializer='he_normal', padding='same', use_bias=False,
kernel_regularizer=l2(weight_decay), name=name_or_none(block_prefix, '_bottleneck_conv2D'))(x)
x = BatchNormalization(axis=concat_axis, epsilon=1.1e-5,
name=name_or_none(block_prefix, '_bottleneck_bn'))(x)
x = Activation('relu')(x)
x = Conv2D(nb_filter, (3, 3), kernel_initializer='he_normal', padding='same', use_bias=False,
name=name_or_none(block_prefix, '_conv2D'))(x)
if dropout_rate:
x = Dropout(dropout_rate)(x)
return x
def __dense_block(x, nb_layers, nb_filter, growth_rate, bottleneck=False, dropout_rate=None,
weight_decay=1e-4, grow_nb_filters=True, return_concat_list=False, block_prefix=None):
'''
Build a dense_block where the output of each conv_block is fed
to subsequent ones
# Arguments
x: input keras tensor
nb_layers: the number of conv_blocks to append to the model
nb_filter: integer, the dimensionality of the output space
(i.e. the number output of filters in the convolution)
growth_rate: growth rate of the dense block
bottleneck: if True, adds a bottleneck convolution block to
each conv_block
dropout_rate: dropout rate
weight_decay: weight decay factor
grow_nb_filters: if True, allows number of filters to grow
return_concat_list: set to True to return the list of
feature maps along with the actual output
block_prefix: str, for block unique naming
# Return
If return_concat_list is True, returns a list of the output
keras tensor, the number of filters and a list of all the
dense blocks added to the keras tensor
If return_concat_list is False, returns a list of the output
keras tensor and the number of filters
'''
with K.name_scope('DenseBlock'):
concat_axis = 1 if K.image_data_format() == 'channels_first' else -1
x_list = [x]
for i in range(nb_layers):
cb = __conv_block(x, growth_rate, bottleneck, dropout_rate, weight_decay,
block_prefix=name_or_none(block_prefix, '_%i' % i))
x_list.append(cb)
x = concatenate([x, cb], axis=concat_axis)
if grow_nb_filters:
nb_filter += growth_rate
if return_concat_list:
return x, nb_filter, x_list
else:
return x, nb_filter
def __transition_block(ip, nb_filter, compression=1.0, weight_decay=1e-4, block_prefix=None, transition_pooling='max'):
'''
Adds a pointwise convolution layer (with batch normalization and relu),
and an average pooling layer. The number of output convolution filters
can be reduced by appropriately reducing the compression parameter.
# Arguments
ip: input keras tensor
nb_filter: integer, the dimensionality of the output space
(i.e. the number output of filters in the convolution)
compression: calculated as 1 - reduction. Reduces the number
of feature maps in the transition block.
weight_decay: weight decay factor
block_prefix: str, for block unique naming
# Input shape
4D tensor with shape:
`(samples, channels, rows, cols)` if data_format='channels_first'
or 4D tensor with shape:
`(samples, rows, cols, channels)` if data_format='channels_last'.
# Output shape
4D tensor with shape:
`(samples, nb_filter * compression, rows / 2, cols / 2)`
if data_format='channels_first'
or 4D tensor with shape:
`(samples, rows / 2, cols / 2, nb_filter * compression)`
if data_format='channels_last'.
# Returns
a keras tensor
'''
with K.name_scope('Transition'):
concat_axis = 1 if K.image_data_format() == 'channels_first' else -1
x = BatchNormalization(axis=concat_axis, epsilon=1.1e-5, name=name_or_none(block_prefix, '_bn'))(ip)
x = Activation('relu')(x)
x = Conv2D(int(nb_filter * compression), (1, 1), kernel_initializer='he_normal', padding='same',
use_bias=False, kernel_regularizer=l2(weight_decay), name=name_or_none(block_prefix, '_conv2D'))(x)
if transition_pooling == 'avg':
x = AveragePooling2D((2, 2), strides=(2, 2))(x)
elif transition_pooling == 'max':
x = MaxPooling2D((2, 2), strides=(2, 2))(x)
return x
def __transition_up_block(ip, nb_filters, type='deconv', weight_decay=1E-4, block_prefix=None):
'''Adds an upsampling block. Upsampling operation relies on the the type parameter.
# Arguments
ip: input keras tensor
nb_filters: integer, the dimensionality of the output space
(i.e. the number output of filters in the convolution)
type: can be 'upsampling', 'subpixel', 'deconv'. Determines
type of upsampling performed
weight_decay: weight decay factor
block_prefix: str, for block unique naming
# Input shape
4D tensor with shape:
`(samples, channels, rows, cols)` if data_format='channels_first'
or 4D tensor with shape:
`(samples, rows, cols, channels)` if data_format='channels_last'.
# Output shape
4D tensor with shape:
`(samples, nb_filter, rows * 2, cols * 2)` if data_format='channels_first'
or 4D tensor with shape:
`(samples, rows * 2, cols * 2, nb_filter)` if data_format='channels_last'.
# Returns
a keras tensor
'''
with K.name_scope('TransitionUp'):
if type == 'upsampling':
x = UpSampling2D(name=name_or_none(block_prefix, '_upsampling'))(ip)
elif type == 'subpixel':
x = Conv2D(nb_filters, (3, 3), activation='relu', padding='same', kernel_regularizer=l2(weight_decay),
use_bias=False, kernel_initializer='he_normal', name=name_or_none(block_prefix, '_conv2D'))(ip)
x = SubPixelUpscaling(scale_factor=2, name=name_or_none(block_prefix, '_subpixel'))(x)
x = Conv2D(nb_filters, (3, 3), activation='relu', padding='same', kernel_regularizer=l2(weight_decay),
use_bias=False, kernel_initializer='he_normal', name=name_or_none(block_prefix, '_conv2D'))(x)
else:
x = Conv2DTranspose(nb_filters, (3, 3), activation='relu', padding='same', strides=(2, 2),
kernel_initializer='he_normal', kernel_regularizer=l2(weight_decay),
name=name_or_none(block_prefix, '_conv2DT'))(ip)
return x
def __create_dense_net(nb_classes, img_input, include_top, depth=40, nb_dense_block=3, growth_rate=12, nb_filter=-1,
nb_layers_per_block=-1, bottleneck=False, reduction=0.0, dropout_rate=None, weight_decay=1e-4,
subsample_initial_block=False, pooling=None, activation='sigmoid', transition_pooling='avg'):
''' Build the DenseNet model
# Arguments
nb_classes: number of classes
img_input: tuple of shape (channels, rows, columns) or (rows, columns, channels)
include_top: flag to include the final Dense layer
depth: number or layers
nb_dense_block: number of dense blocks to add to end (generally = 3)
growth_rate: number of filters to add per dense block
nb_filter: initial number of filters. Default -1 indicates initial number of filters is 2 * growth_rate
nb_layers_per_block: number of layers in each dense block.
Can be a -1, positive integer or a list.
If -1, calculates nb_layer_per_block from the depth of the network.
If positive integer, a set number of layers per dense block.
If list, nb_layer is used as provided. Note that list size must
be (nb_dense_block + 1)
bottleneck: add bottleneck blocks
reduction: reduction factor of transition blocks. Note : reduction value is inverted to compute compression
dropout_rate: dropout rate
weight_decay: weight decay rate
subsample_initial_block: Changes model type to suit different datasets.
Should be set to True for ImageNet, and False for CIFAR datasets.
When set to True, the initial convolution will be strided and
adds a MaxPooling2D before the initial dense block.
pooling: Optional pooling mode for feature extraction
when `include_top` is `False`.
- `None` means that the output of the model
will be the 4D tensor output of the
last convolutional layer.
- `avg` means that global average pooling
will be applied to the output of the
last convolutional layer, and thus
the output of the model will be a
2D tensor.
- `max` means that global max pooling will
be applied.
activation: Type of activation at the top layer. Can be one of 'softmax' or 'sigmoid'.
Note that if sigmoid is used, classes must be 1.
transition_pooling: `avg` for avg pooling (default), `max` for max pooling,
None for no pooling during scale transition blocks. Please note that this
default differs from the DenseNetFCN paper in accordance with the DenseNet
paper.
# Returns
a keras tensor
# Raises
ValueError: in case of invalid argument for `reduction`
or `nb_dense_block`
'''
with K.name_scope('DenseNet'):
concat_axis = 1 if K.image_data_format() == 'channels_first' else -1
if reduction != 0.0:
if not (reduction <= 1.0 and reduction > 0.0):
raise ValueError('`reduction` value must lie between 0.0 and 1.0')
# layers in each dense block
if type(nb_layers_per_block) is list or type(nb_layers_per_block) is tuple:
nb_layers = list(nb_layers_per_block) # Convert tuple to list
if len(nb_layers) != (nb_dense_block):
raise ValueError('If `nb_dense_block` is a list, its length must match '
'the number of layers provided by `nb_layers`.')
final_nb_layer = nb_layers[-1]
nb_layers = nb_layers[:-1]
else:
if nb_layers_per_block == -1:
assert (depth - 4) % 3 == 0, 'Depth must be 3 N + 4 if nb_layers_per_block == -1'
count = int((depth - 4) / 3)
if bottleneck:
count = count // 2
nb_layers = [count for _ in range(nb_dense_block)]
final_nb_layer = count
else:
final_nb_layer = nb_layers_per_block
nb_layers = [nb_layers_per_block] * nb_dense_block
# compute initial nb_filter if -1, else accept users initial nb_filter
if nb_filter <= 0:
nb_filter = 2 * growth_rate
# compute compression factor
compression = 1.0 - reduction
# Initial convolution
if subsample_initial_block:
initial_kernel = (7, 7)
initial_strides = (2, 2)
else:
initial_kernel = (3, 3)
initial_strides = (1, 1)
x = Conv2D(nb_filter, initial_kernel, kernel_initializer='he_normal', padding='same', name='initial_conv2D',
strides=initial_strides, use_bias=False, kernel_regularizer=l2(weight_decay))(img_input)
if subsample_initial_block:
x = BatchNormalization(axis=concat_axis, epsilon=1.1e-5, name='initial_bn')(x)
x = Activation('relu')(x)
x = MaxPooling2D((3, 3), strides=(2, 2), padding='same')(x)
# Add dense blocks
for block_idx in range(nb_dense_block - 1):
x, nb_filter = __dense_block(x, nb_layers[block_idx], nb_filter, growth_rate, bottleneck=bottleneck,
dropout_rate=dropout_rate, weight_decay=weight_decay,
block_prefix='dense_%i' % block_idx)
# add transition_block
x = __transition_block(x, nb_filter, compression=compression, weight_decay=weight_decay,
block_prefix='tr_%i' % block_idx, transition_pooling=transition_pooling)
nb_filter = int(nb_filter * compression)
# The last dense_block does not have a transition_block
x, nb_filter = __dense_block(x, final_nb_layer, nb_filter, growth_rate, bottleneck=bottleneck,
dropout_rate=dropout_rate, weight_decay=weight_decay,
block_prefix='dense_%i' % (nb_dense_block - 1))
x = BatchNormalization(axis=concat_axis, epsilon=1.1e-5, name='final_bn')(x)
x = Activation('relu')(x)
if include_top:
if pooling == 'avg':
x = GlobalAveragePooling2D()(x)
elif pooling == 'max':
x = GlobalMaxPooling2D()(x)
x = Dense(nb_classes, activation=activation)(x)
else:
if pooling == 'avg':
x = GlobalAveragePooling2D()(x)
elif pooling == 'max':
x = GlobalMaxPooling2D()(x)
return x
def __create_fcn_dense_net(nb_classes, img_input, include_top, nb_dense_block=5, growth_rate=12,
reduction=0.0, dropout_rate=None, weight_decay=1e-4,
nb_layers_per_block=4, nb_upsampling_conv=128, upsampling_type='deconv',
init_conv_filters=48, input_shape=None, activation='sigmoid',
early_transition=False, transition_pooling='max', initial_kernel_size=(3, 3)):
''' Build the DenseNet-FCN model
# Arguments
nb_classes: number of classes
img_input: tuple of shape (channels, rows, columns) or (rows, columns, channels)
include_top: flag to include the final Dense layer
nb_dense_block: number of dense blocks to add to end (generally = 3)
growth_rate: number of filters to add per dense block
reduction: reduction factor of transition blocks. Note : reduction value is inverted to compute compression
dropout_rate: dropout rate
weight_decay: weight decay
nb_layers_per_block: number of layers in each dense block.
Can be a positive integer or a list.
If positive integer, a set number of layers per dense block.
If list, nb_layer is used as provided. Note that list size must
be (nb_dense_block + 1)
nb_upsampling_conv: number of convolutional layers in upsampling via subpixel convolution
upsampling_type: Can be one of 'upsampling', 'deconv' and 'subpixel'. Defines
type of upsampling algorithm used.
input_shape: Only used for shape inference in fully convolutional networks.
activation: Type of activation at the top layer. Can be one of 'softmax' or 'sigmoid'.
Note that if sigmoid is used, classes must be 1.
early_transition: Start with an extra initial transition down and end with an extra
transition up to reduce the network size.
transition_pooling: 'max' for max pooling (default), 'avg' for average pooling,
None for no pooling. Please note that this default differs from the DenseNet
paper in accordance with the DenseNetFCN paper.
initial_kernel_size: The first Conv2D kernel might vary in size based on the
application, this parameter makes it configurable.
# Returns
a keras tensor
# Raises
ValueError: in case of invalid argument for `reduction`,
`nb_dense_block` or `nb_upsampling_conv`.
'''
with K.name_scope('DenseNetFCN'):
concat_axis = 1 if K.image_data_format() == 'channels_first' else -1
if concat_axis == 1: # channels_first dim ordering
_, rows, cols = input_shape
else:
rows, cols, _ = input_shape
if reduction != 0.0:
if not (reduction <= 1.0 and reduction > 0.0):
raise ValueError('`reduction` value must lie between 0.0 and 1.0')
# check if upsampling_conv has minimum number of filters
# minimum is set to 12, as at least 3 color channels are needed for correct upsampling
if not (nb_upsampling_conv > 12 and nb_upsampling_conv % 4 == 0):
raise ValueError('Parameter `nb_upsampling_conv` number of channels must '
'be a positive number divisible by 4 and greater than 12')
# layers in each dense block
if type(nb_layers_per_block) is list or type(nb_layers_per_block) is tuple:
nb_layers = list(nb_layers_per_block) # Convert tuple to list
if len(nb_layers) != (nb_dense_block + 1):
raise ValueError('If `nb_dense_block` is a list, its length must be '
'(`nb_dense_block` + 1)')
bottleneck_nb_layers = nb_layers[-1]
rev_layers = nb_layers[::-1]
nb_layers.extend(rev_layers[1:])
else:
bottleneck_nb_layers = nb_layers_per_block
nb_layers = [nb_layers_per_block] * (2 * nb_dense_block + 1)
# compute compression factor
compression = 1.0 - reduction
# Initial convolution
x = Conv2D(init_conv_filters, initial_kernel_size, kernel_initializer='he_normal', padding='same', name='initial_conv2D',
use_bias=False, kernel_regularizer=l2(weight_decay))(img_input)
x = BatchNormalization(axis=concat_axis, epsilon=1.1e-5, name='initial_bn')(x)
x = Activation('relu')(x)
nb_filter = init_conv_filters
skip_list = []
if early_transition:
x = __transition_block(x, nb_filter, compression=compression, weight_decay=weight_decay,
block_prefix='tr_early', transition_pooling=transition_pooling)
# Add dense blocks and transition down block
for block_idx in range(nb_dense_block):
x, nb_filter = __dense_block(x, nb_layers[block_idx], nb_filter, growth_rate, dropout_rate=dropout_rate,
weight_decay=weight_decay, block_prefix='dense_%i' % block_idx)
# Skip connection
skip_list.append(x)
# add transition_block
x = __transition_block(x, nb_filter, compression=compression, weight_decay=weight_decay,
block_prefix='tr_%i' % block_idx, transition_pooling=transition_pooling)
nb_filter = int(nb_filter * compression) # this is calculated inside transition_down_block
# The last dense_block does not have a transition_down_block
# return the concatenated feature maps without the concatenation of the input
_, nb_filter, concat_list = __dense_block(x, bottleneck_nb_layers, nb_filter, growth_rate,
dropout_rate=dropout_rate, weight_decay=weight_decay,
return_concat_list=True,
block_prefix='dense_%i' % nb_dense_block)
skip_list = skip_list[::-1] # reverse the skip list
# Add dense blocks and transition up block
for block_idx in range(nb_dense_block):
n_filters_keep = growth_rate * nb_layers[nb_dense_block + block_idx]
# upsampling block must upsample only the feature maps (concat_list[1:]),
# not the concatenation of the input with the feature maps (concat_list[0].
l = concatenate(concat_list[1:], axis=concat_axis)
t = __transition_up_block(l, nb_filters=n_filters_keep, type=upsampling_type, weight_decay=weight_decay,
block_prefix='tr_up_%i' % block_idx)
# concatenate the skip connection with the transition block
x = concatenate([t, skip_list[block_idx]], axis=concat_axis)
# Dont allow the feature map size to grow in upsampling dense blocks
x_up, nb_filter, concat_list = __dense_block(x, nb_layers[nb_dense_block + block_idx + 1],
nb_filter=growth_rate, growth_rate=growth_rate,
dropout_rate=dropout_rate, weight_decay=weight_decay,
return_concat_list=True, grow_nb_filters=False,
block_prefix='dense_%i' % (nb_dense_block + 1 + block_idx))
if early_transition:
x_up = __transition_up_block(x_up, nb_filters=nb_filter, type=upsampling_type, weight_decay=weight_decay,
block_prefix='tr_up_early')
if include_top:
x = Conv2D(nb_classes, (1, 1), activation='linear', padding='same', use_bias=False)(x_up)
if K.image_data_format() == 'channels_first':
channel, row, col = input_shape
else:
row, col, channel = input_shape
x = Reshape((row * col, nb_classes))(x)
x = Activation(activation)(x)
x = Reshape((row, col, nb_classes))(x)
else:
x = x_up
return x
================================================
FILE: lib/extract_patches.py
================================================
import numpy as np
import random
import configparser
from .help_functions import load_hdf5
from .help_functions import visualize
from .help_functions import group_images
from .pre_processing import my_PreProc
#To select the same images
# random.seed(10)
#Load the original data and return the extracted patches for training/testing
def get_data_training(DRIVE_train_imgs_original,
DRIVE_train_groudTruth,
patch_height,
patch_width,
N_subimgs,
inside_FOV):
train_imgs_original = load_hdf5(DRIVE_train_imgs_original)
train_masks = load_hdf5(DRIVE_train_groudTruth) #masks always the same
# visualize(group_images(train_imgs_original[0:20,:,:,:],5),'imgs_train')#.show() #check original imgs train
train_imgs = my_PreProc(train_imgs_original)
train_masks = train_masks/255.
train_imgs = train_imgs[:,:,9:574,:] #cut bottom and top so now it is 565*565
train_masks = train_masks[:,:,9:574,:] #cut bottom and top so now it is 565*565
data_consistency_check(train_imgs,train_masks)
#check masks are within 0-1
assert(np.min(train_masks)==0 and np.max(train_masks)==1)
print("\ntrain images/masks shape:")
print(train_imgs.shape)
print("train images range (min-max): " +str(np.min(train_imgs)) +' - '+str(np.max(train_imgs)))
print("train masks are within 0-1\n")
#extract the TRAINING patches from the full images
patches_imgs_train, patches_masks_train = extract_random(train_imgs,train_masks,patch_height,patch_width,N_subimgs,inside_FOV)
data_consistency_check(patches_imgs_train, patches_masks_train)
print("\ntrain PATCHES images/masks shape:")
print(patches_imgs_train.shape)
print("train PATCHES images range (min-max): " +str(np.min(patches_imgs_train)) +' - '+str(np.max(patches_imgs_train)))
return patches_imgs_train, patches_masks_train#, patches_imgs_test, patches_masks_test
#Load the original data and return the extracted patches for training/testing
def get_data_testing(DRIVE_test_imgs_original, DRIVE_test_groudTruth, Imgs_to_test, patch_height, patch_width):
### test
test_imgs_original = load_hdf5(DRIVE_test_imgs_original)
test_masks = load_hdf5(DRIVE_test_groudTruth)
test_imgs = my_PreProc(test_imgs_original)
test_masks = test_masks/255.
#extend both images and masks so they can be divided exactly by the patches dimensions
test_imgs = test_imgs[0:Imgs_to_test,:,:,:]
test_masks = test_masks[0:Imgs_to_test,:,:,:]
test_imgs = paint_border(test_imgs,patch_height,patch_width)
test_masks = paint_border(test_masks,patch_height,patch_width)
data_consistency_check(test_imgs, test_masks)
#check masks are within 0-1
assert(np.max(test_masks)==1 and np.min(test_masks)==0)
print("\ntest images/masks shape:")
print(test_imgs.shape)
print("test images range (min-max): " +str(np.min(test_imgs)) +' - '+str(np.max(test_imgs)))
print("test masks are within 0-1\n")
#extract the TEST patches from the full images
patches_imgs_test = extract_ordered(test_imgs,patch_height,patch_width)
patches_masks_test = extract_ordered(test_masks,patch_height,patch_width)
data_consistency_check(patches_imgs_test, patches_masks_test)
print("\ntest PATCHES images/masks shape:")
print(patches_imgs_test.shape)
print("test PATCHES images range (min-max): " +str(np.min(patches_imgs_test)) +' - '+str(np.max(patches_imgs_test)))
return patches_imgs_test, patches_masks_test
# Load the original data and return the extracted patches for testing
# return the ground truth in its original shape
def get_data_testing_overlap(DRIVE_test_imgs_original, DRIVE_test_groudTruth, Imgs_to_test, patch_height, patch_width, stride_height, stride_width):
### test
test_imgs_original = load_hdf5(DRIVE_test_imgs_original)
test_masks = load_hdf5(DRIVE_test_groudTruth)
test_imgs = my_PreProc(test_imgs_original)
test_masks = test_masks/255.
#extend both images and masks so they can be divided exactly by the patches dimensions
test_imgs = test_imgs[0:Imgs_to_test,:,:,:]
test_masks = test_masks[0:Imgs_to_test,:,:,:]
test_imgs = paint_border_overlap(test_imgs, patch_height, patch_width, stride_height, stride_width)
#check masks are within 0-1
assert(np.max(test_masks)==1 and np.min(test_masks)==0)
print("\ntest images shape:")
print(test_imgs.shape)
print("\ntest mask shape:")
print(test_masks.shape)
print("test images range (min-max): " +str(np.min(test_imgs)) +' - '+str(np.max(test_imgs)))
print("test masks are within 0-1\n")
#extract the TEST patches from the full images
patches_imgs_test = extract_ordered_overlap(test_imgs,patch_height,patch_width,stride_height,stride_width)
print("\ntest PATCHES images shape:")
print(patches_imgs_test.shape)
print("test PATCHES images range (min-max): " +str(np.min(patches_imgs_test)) +' - '+str(np.max(patches_imgs_test)))
return patches_imgs_test, test_imgs.shape[2], test_imgs.shape[3], test_masks
#data consinstency check
def data_consistency_check(imgs,masks):
assert(len(imgs.shape)==len(masks.shape))
assert(imgs.shape[0]==masks.shape[0])
assert(imgs.shape[2]==masks.shape[2])
assert(imgs.shape[3]==masks.shape[3])
assert(masks.shape[1]==1)
assert(imgs.shape[1]==1 or imgs.shape[1]==3)
#extract patches randomly in the full training images
# -- Inside OR in full image
def extract_random(full_imgs,full_masks, patch_h,patch_w, N_patches, inside=True):
if (N_patches%full_imgs.shape[0] != 0):
print("N_patches: plase enter a multiple of 20")
exit()
assert (len(full_imgs.shape)==4 and len(full_masks.shape)==4) #4D arrays
assert (full_imgs.shape[1]==1 or full_imgs.shape[1]==3) #check the channel is 1 or 3
assert (full_masks.shape[1]==1) #masks only black and white
assert (full_imgs.shape[2] == full_masks.shape[2] and full_imgs.shape[3] == full_masks.shape[3])
patches = np.empty((N_patches,full_imgs.shape[1],patch_h,patch_w))
patches_masks = np.empty((N_patches,full_masks.shape[1],patch_h,patch_w))
img_h = full_imgs.shape[2] #height of the full image
img_w = full_imgs.shape[3] #width of the full image
# (0,0) in the center of the image
patch_per_img = int(N_patches/full_imgs.shape[0]) #N_patches equally divided in the full images
print("patches per full image: " +str(patch_per_img))
iter_tot = 0 #iter over the total numbe rof patches (N_patches)
for i in range(full_imgs.shape[0]): #loop over the full images
k=0
while k <patch_per_img:
x_center = random.randint(0+int(patch_w/2),img_w-int(patch_w/2))
# print "x_center " +str(x_center)
y_center = random.randint(0+int(patch_h/2),img_h-int(patch_h/2))
# print "y_center " +str(y_center)
#check whether the patch is fully contained in the FOV
if inside==True:
if is_patch_inside_FOV(x_center,y_center,img_w,img_h,patch_h)==False:
continue
patch = full_imgs[i,:,y_center-int(patch_h/2):y_center+int(patch_h/2),x_center-int(patch_w/2):x_center+int(patch_w/2)]
patch_mask = full_masks[i,:,y_center-int(patch_h/2):y_center+int(patch_h/2),x_center-int(patch_w/2):x_center+int(patch_w/2)]
patches[iter_tot]=patch
patches_masks[iter_tot]=patch_mask
iter_tot +=1 #total
k+=1 #per full_img
return patches, patches_masks
#check if the patch is fully contained in the FOV
def is_patch_inside_FOV(x,y,img_w,img_h,patch_h):
x_ = x - int(img_w/2) # origin (0,0) shifted to image center
y_ = y - int(img_h/2) # origin (0,0) shifted to image center
R_inside = 270 - int(patch_h * np.sqrt(2.0) / 2.0) #radius is 270 (from DRIVE db docs), minus the patch diagonal (assumed it is a square #this is the limit to contain the full patch in the FOV
radius = np.sqrt((x_*x_)+(y_*y_))
if radius < R_inside:
return True
else:
return False
#Divide all the full_imgs in pacthes
def extract_ordered(full_imgs, patch_h, patch_w):
assert (len(full_imgs.shape)==4) #4D arrays
assert (full_imgs.shape[1]==1 or full_imgs.shape[1]==3) #check the channel is 1 or 3
img_h = full_imgs.shape[2] #height of the full image
img_w = full_imgs.shape[3] #width of the full image
N_patches_h = int(img_h/patch_h) #round to lowest int
if (img_h%patch_h != 0):
print("warning: " +str(N_patches_h) +" patches in height, with about " +str(img_h%patch_h) +" pixels left over")
N_patches_w = int(img_w/patch_w) #round to lowest int
if (img_h%patch_h != 0):
print("warning: " +str(N_patches_w) +" patches in width, with about " +str(img_w%patch_w) +" pixels left over")
print("number of patches per image: " +str(N_patches_h*N_patches_w))
N_patches_tot = (N_patches_h*N_patches_w)*full_imgs.shape[0]
patches = np.empty((N_patches_tot,full_imgs.shape[1],patch_h,patch_w))
iter_tot = 0 #iter over the total number of patches (N_patches)
for i in range(full_imgs.shape[0]): #loop over the full images
for h in range(N_patches_h):
for w in range(N_patches_w):
patch = full_imgs[i,:,h*patch_h:(h*patch_h)+patch_h,w*patch_w:(w*patch_w)+patch_w]
patches[iter_tot]=patch
iter_tot +=1 #total
assert (iter_tot==N_patches_tot)
return patches #array with all the full_imgs divided in patches
def paint_border_overlap(full_imgs, patch_h, patch_w, stride_h, stride_w):
assert (len(full_imgs.shape)==4) #4D arrays
assert (full_imgs.shape[1]==1 or full_imgs.shape[1]==3) #check the channel is 1 or 3
img_h = full_imgs.shape[2] #height of the full image
img_w = full_imgs.shape[3] #width of the full image
leftover_h = (img_h-patch_h)%stride_h #leftover on the h dim
leftover_w = (img_w-patch_w)%stride_w #leftover on the w dim
if (leftover_h != 0): #change dimension of img_h
print("\nthe side H is not compatible with the selected stride of " +str(stride_h))
print("img_h " +str(img_h) + ", patch_h " +str(patch_h) + ", stride_h " +str(stride_h))
print("(img_h - patch_h) MOD stride_h: " +str(leftover_h))
print("So the H dim will be padded with additional " +str(stride_h - leftover_h) + " pixels")
tmp_full_imgs = np.zeros((full_imgs.shape[0],full_imgs.shape[1],img_h+(stride_h-leftover_h),img_w))
tmp_full_imgs[0:full_imgs.shape[0],0:full_imgs.shape[1],0:img_h,0:img_w] = full_imgs
full_imgs = tmp_full_imgs
if (leftover_w != 0): #change dimension of img_w
print("the side W is not compatible with the selected stride of " +str(stride_w))
print("img_w " +str(img_w) + ", patch_w " +str(patch_w) + ", stride_w " +str(stride_w))
print("(img_w - patch_w) MOD stride_w: " +str(leftover_w))
print("So the W dim will be padded with additional " +str(stride_w - leftover_w) + " pixels")
tmp_full_imgs = np.zeros((full_imgs.shape[0],full_imgs.shape[1],full_imgs.shape[2],img_w+(stride_w - leftover_w)))
tmp_full_imgs[0:full_imgs.shape[0],0:full_imgs.shape[1],0:full_imgs.shape[2],0:img_w] = full_imgs
full_imgs = tmp_full_imgs
print("new full images shape: \n" +str(full_imgs.shape))
return full_imgs
#Divide all the full_imgs in pacthes
def extract_ordered_overlap(full_imgs, patch_h, patch_w,stride_h,stride_w):
assert (len(full_imgs.shape)==4) #4D arrays
assert (full_imgs.shape[1]==1 or full_imgs.shape[1]==3) #check the channel is 1 or 3
img_h = full_imgs.shape[2] #height of the full image
img_w = full_imgs.shape[3] #width of the full image
assert ((img_h-patch_h)%stride_h==0 and (img_w-patch_w)%stride_w==0)
N_patches_img = ((img_h-patch_h)//stride_h+1)*((img_w-patch_w)//stride_w+1) #// --> division between integers
N_patches_tot = N_patches_img*full_imgs.shape[0]
print("Number of patches on h : " +str(((img_h-patch_h)//stride_h+1)))
print("Number of patches on w : " +str(((img_w-patch_w)//stride_w+1)))
print("number of patches per image: " +str(N_patches_img) +", totally for this dataset: " +str(N_patches_tot))
patches = np.empty((N_patches_tot,full_imgs.shape[1],patch_h,patch_w))
iter_tot = 0 #iter over the total number of patches (N_patches)
for i in range(full_imgs.shape[0]): #loop over the full images
for h in range((img_h-patch_h)//stride_h+1):
for w in range((img_w-patch_w)//stride_w+1):
patch = full_imgs[i,:,h*stride_h:(h*stride_h)+patch_h,w*stride_w:(w*stride_w)+patch_w]
patches[iter_tot]=patch
iter_tot +=1 #total
assert (iter_tot==N_patches_tot)
return patches #array with all the full_imgs divided in patches
def recompone_overlap(preds, img_h, img_w, stride_h, stride_w):
assert (len(preds.shape)==4) #4D arrays
assert (preds.shape[1]==1 or preds.shape[1]==3) #check the channel is 1 or 3
patch_h = preds.shape[2]
patch_w = preds.shape[3]
N_patches_h = (img_h-patch_h)//stride_h+1
N_patches_w = (img_w-patch_w)//stride_w+1
N_patches_img = N_patches_h * N_patches_w
print("N_patches_h: " +str(N_patches_h))
print("N_patches_w: " +str(N_patches_w))
print("N_patches_img: " +str(N_patches_img))
assert (preds.shape[0]%N_patches_img==0)
N_full_imgs = preds.shape[0]//N_patches_img
print("According to the dimension inserted, there are " +str(N_full_imgs) +" full images (of " +str(img_h)+"x" +str(img_w) +" each)")
full_prob = np.zeros((N_full_imgs,preds.shape[1],img_h,img_w)) #itialize to zero mega array with sum of Probabilities
full_sum = np.zeros((N_full_imgs,preds.shape[1],img_h,img_w))
k = 0 #iterator over all the patches
for i in range(N_full_imgs):
for h in range((img_h-patch_h)//stride_h+1):
for w in range((img_w-patch_w)//stride_w+1):
full_prob[i,:,h*stride_h:(h*stride_h)+patch_h,w*stride_w:(w*stride_w)+patch_w]+=preds[k]
full_sum[i,:,h*stride_h:(h*stride_h)+patch_h,w*stride_w:(w*stride_w)+patch_w]+=1
k+=1
assert(k==preds.shape[0])
assert(np.min(full_sum)>=1.0) #at least one
final_avg = full_prob/full_sum
print(final_avg.shape)
assert(np.max(final_avg)<=1.0) #max value for a pixel is 1.0
assert(np.min(final_avg)>=0.0) #min value for a pixel is 0.0
return final_avg
#Recompone the full images with the patches
def recompone(data,N_h,N_w):
assert (data.shape[1]==1 or data.shape[1]==3) #check the channel is 1 or 3
assert(len(data.shape)==4)
N_pacth_per_img = N_w*N_h
assert(data.shape[0]%N_pacth_per_img == 0)
N_full_imgs = data.shape[0]/N_pacth_per_img
patch_h = data.shape[2]
patch_w = data.shape[3]
N_pacth_per_img = N_w*N_h
#define and start full recompone
full_recomp = np.empty((N_full_imgs,data.shape[1],N_h*patch_h,N_w*patch_w))
k = 0 #iter full img
s = 0 #iter single patch
while (s<data.shape[0]):
#recompone one:
single_recon = np.empty((data.shape[1],N_h*patch_h,N_w*patch_w))
for h in range(N_h):
for w in range(N_w):
single_recon[:,h*patch_h:(h*patch_h)+patch_h,w*patch_w:(w*patch_w)+patch_w]=data[s]
s+=1
full_recomp[k]=single_recon
k+=1
assert (k==N_full_imgs)
return full_recomp
#Extend the full images becasue patch divison is not exact
def paint_border(data,patch_h,patch_w):
assert (len(data.shape)==4) #4D arrays
assert (data.shape[1]==1 or data.shape[1]==3) #check the channel is 1 or 3
img_h=data.shape[2]
img_w=data.shape[3]
new_img_h = 0
new_img_w = 0
if (img_h%patch_h)==0:
new_img_h = img_h
else:
new_img_h = ((int(img_h)/int(patch_h))+1)*patch_h
if (img_w%patch_w)==0:
new_img_w = img_w
else:
new_img_w = ((int(img_w)/int(patch_w))+1)*patch_w
new_data = np.zeros((data.shape[0],data.shape[1],new_img_h,new_img_w))
new_data[:,:,0:img_h,0:img_w] = data[:,:,:,:]
return new_data
#return only the pixels contained in the FOV, for both images and masks
def pred_only_FOV(data_imgs,data_masks,original_imgs_border_masks):
assert (len(data_imgs.shape)==4 and len(data_masks.shape)==4) #4D arrays
assert (data_imgs.shape[0]==data_masks.shape[0])
assert (data_imgs.shape[2]==data_masks.shape[2])
assert (data_imgs.shape[3]==data_masks.shape[3])
assert (data_imgs.shape[1]==1 and data_masks.shape[1]==1) #check the channel is 1
height = data_imgs.shape[2]
width = data_imgs.shape[3]
new_pred_imgs = []
new_pred_masks = []
for i in range(data_imgs.shape[0]): #loop over the full images
for x in range(width):
for y in range(height):
if inside_FOV_DRIVE(i,x,y,original_imgs_border_masks)==True:
new_pred_imgs.append(data_imgs[i,:,y,x])
new_pred_masks.append(data_masks[i,:,y,x])
new_pred_imgs = np.asarray(new_pred_imgs)
new_pred_masks = np.asarray(new_pred_masks)
return new_pred_imgs, new_pred_masks
#function to set to black everything outside the FOV, in a full image
def kill_border(data, original_imgs_border_masks):
assert (len(data.shape)==4) #4D arrays
assert (data.shape[1]==1 or data.shape[1]==3) #check the channel is 1 or 3
height = data.shape[2]
width = data.shape[3]
for i in range(data.shape[0]): #loop over the full images
for x in range(width):
for y in range(height):
if inside_FOV_DRIVE(i,x,y,original_imgs_border_masks)==False:
data[i,:,y,x]=0.0
def inside_FOV_DRIVE(i, x, y, DRIVE_masks):
assert (len(DRIVE_masks.shape)==4) #4D arrays
assert (DRIVE_masks.shape[1]==1) #DRIVE masks is black and white
# DRIVE_masks = DRIVE_masks/255. #NOOO!! otherwise with float numbers takes forever!!
if (x >= DRIVE_masks.shape[3] or y >= DRIVE_masks.shape[2]): #my image bigger than the original
return False
if (DRIVE_masks[i,0,y,x]>0): #0==black pixels
# print DRIVE_masks[i,0,y,x] #verify it is working right
return True
else:
return False
================================================
FILE: lib/extract_patches.py.bak
================================================
import numpy as np
import random
import ConfigParser
from help_functions import load_hdf5
from help_functions import visualize
from help_functions import group_images
from pre_processing import my_PreProc
#To select the same images
# random.seed(10)
#Load the original data and return the extracted patches for training/testing
def get_data_training(DRIVE_train_imgs_original,
DRIVE_train_groudTruth,
patch_height,
patch_width,
N_subimgs,
inside_FOV):
train_imgs_original = load_hdf5(DRIVE_train_imgs_original)
train_masks = load_hdf5(DRIVE_train_groudTruth) #masks always the same
# visualize(group_images(train_imgs_original[0:20,:,:,:],5),'imgs_train')#.show() #check original imgs train
train_imgs = my_PreProc(train_imgs_original)
train_masks = train_masks/255.
train_imgs = train_imgs[:,:,9:574,:] #cut bottom and top so now it is 565*565
train_masks = train_masks[:,:,9:574,:] #cut bottom and top so now it is 565*565
data_consistency_check(train_imgs,train_masks)
#check masks are within 0-1
assert(np.min(train_masks)==0 and np.max(train_masks)==1)
print "\ntrain images/masks shape:"
print train_imgs.shape
print "train images range (min-max): " +str(np.min(train_imgs)) +' - '+str(np.max(train_imgs))
print "train masks are within 0-1\n"
#extract the TRAINING patches from the full images
patches_imgs_train, patches_masks_train = extract_random(train_imgs,train_masks,patch_height,patch_width,N_subimgs,inside_FOV)
data_consistency_check(patches_imgs_train, patches_masks_train)
print "\ntrain PATCHES images/masks shape:"
print patches_imgs_train.shape
print "train PATCHES images range (min-max): " +str(np.min(patches_imgs_train)) +' - '+str(np.max(patches_imgs_train))
return patches_imgs_train, patches_masks_train#, patches_imgs_test, patches_masks_test
#Load the original data and return the extracted patches for training/testing
def get_data_testing(DRIVE_test_imgs_original, DRIVE_test_groudTruth, Imgs_to_test, patch_height, patch_width):
### test
test_imgs_original = load_hdf5(DRIVE_test_imgs_original)
test_masks = load_hdf5(DRIVE_test_groudTruth)
test_imgs = my_PreProc(test_imgs_original)
test_masks = test_masks/255.
#extend both images and masks so they can be divided exactly by the patches dimensions
test_imgs = test_imgs[0:Imgs_to_test,:,:,:]
test_masks = test_masks[0:Imgs_to_test,:,:,:]
test_imgs = paint_border(test_imgs,patch_height,patch_width)
test_masks = paint_border(test_masks,patch_height,patch_width)
data_consistency_check(test_imgs, test_masks)
#check masks are within 0-1
assert(np.max(test_masks)==1 and np.min(test_masks)==0)
print "\ntest images/masks shape:"
print test_imgs.shape
print "test images range (min-max): " +str(np.min(test_imgs)) +' - '+str(np.max(test_imgs))
print "test masks are within 0-1\n"
#extract the TEST patches from the full images
patches_imgs_test = extract_ordered(test_imgs,patch_height,patch_width)
patches_masks_test = extract_ordered(test_masks,patch_height,patch_width)
data_consistency_check(patches_imgs_test, patches_masks_test)
print "\ntest PATCHES images/masks shape:"
print patches_imgs_test.shape
print "test PATCHES images range (min-max): " +str(np.min(patches_imgs_test)) +' - '+str(np.max(patches_imgs_test))
return patches_imgs_test, patches_masks_test
# Load the original data and return the extracted patches for testing
# return the ground truth in its original shape
def get_data_testing_overlap(DRIVE_test_imgs_original, DRIVE_test_groudTruth, Imgs_to_test, patch_height, patch_width, stride_height, stride_width):
### test
test_imgs_original = load_hdf5(DRIVE_test_imgs_original)
test_masks = load_hdf5(DRIVE_test_groudTruth)
test_imgs = my_PreProc(test_imgs_original)
test_masks = test_masks/255.
#extend both images and masks so they can be divided exactly by the patches dimensions
test_imgs = test_imgs[0:Imgs_to_test,:,:,:]
test_masks = test_masks[0:Imgs_to_test,:,:,:]
test_imgs = paint_border_overlap(test_imgs, patch_height, patch_width, stride_height, stride_width)
#check masks are within 0-1
assert(np.max(test_masks)==1 and np.min(test_masks)==0)
print "\ntest images shape:"
print test_imgs.shape
print "\ntest mask shape:"
print test_masks.shape
print "test images range (min-max): " +str(np.min(test_imgs)) +' - '+str(np.max(test_imgs))
print "test masks are within 0-1\n"
#extract the TEST patches from the full images
patches_imgs_test = extract_ordered_overlap(test_imgs,patch_height,patch_width,stride_height,stride_width)
print "\ntest PATCHES images shape:"
print patches_imgs_test.shape
print "test PATCHES images range (min-max): " +str(np.min(patches_imgs_test)) +' - '+str(np.max(patches_imgs_test))
return patches_imgs_test, test_imgs.shape[2], test_imgs.shape[3], test_masks
#data consinstency check
def data_consistency_check(imgs,masks):
assert(len(imgs.shape)==len(masks.shape))
assert(imgs.shape[0]==masks.shape[0])
assert(imgs.shape[2]==masks.shape[2])
assert(imgs.shape[3]==masks.shape[3])
assert(masks.shape[1]==1)
assert(imgs.shape[1]==1 or imgs.shape[1]==3)
#extract patches randomly in the full training images
# -- Inside OR in full image
def extract_random(full_imgs,full_masks, patch_h,patch_w, N_patches, inside=True):
if (N_patches%full_imgs.shape[0] != 0):
print "N_patches: plase enter a multiple of 20"
exit()
assert (len(full_imgs.shape)==4 and len(full_masks.shape)==4) #4D arrays
assert (full_imgs.shape[1]==1 or full_imgs.shape[1]==3) #check the channel is 1 or 3
assert (full_masks.shape[1]==1) #masks only black and white
assert (full_imgs.shape[2] == full_masks.shape[2] and full_imgs.shape[3] == full_masks.shape[3])
patches = np.empty((N_patches,full_imgs.shape[1],patch_h,patch_w))
patches_masks = np.empty((N_patches,full_masks.shape[1],patch_h,patch_w))
img_h = full_imgs.shape[2] #height of the full image
img_w = full_imgs.shape[3] #width of the full image
# (0,0) in the center of the image
patch_per_img = int(N_patches/full_imgs.shape[0]) #N_patches equally divided in the full images
print "patches per full image: " +str(patch_per_img)
iter_tot = 0 #iter over the total numbe rof patches (N_patches)
for i in range(full_imgs.shape[0]): #loop over the full images
k=0
while k <patch_per_img:
x_center = random.randint(0+int(patch_w/2),img_w-int(patch_w/2))
# print "x_center " +str(x_center)
y_center = random.randint(0+int(patch_h/2),img_h-int(patch_h/2))
# print "y_center " +str(y_center)
#check whether the patch is fully contained in the FOV
if inside==True:
if is_patch_inside_FOV(x_center,y_center,img_w,img_h,patch_h)==False:
continue
patch = full_imgs[i,:,y_center-int(patch_h/2):y_center+int(patch_h/2),x_center-int(patch_w/2):x_center+int(patch_w/2)]
patch_mask = full_masks[i,:,y_center-int(patch_h/2):y_center+int(patch_h/2),x_center-int(patch_w/2):x_center+int(patch_w/2)]
patches[iter_tot]=patch
patches_masks[iter_tot]=patch_mask
iter_tot +=1 #total
k+=1 #per full_img
return patches, patches_masks
#check if the patch is fully contained in the FOV
def is_patch_inside_FOV(x,y,img_w,img_h,patch_h):
x_ = x - int(img_w/2) # origin (0,0) shifted to image center
y_ = y - int(img_h/2) # origin (0,0) shifted to image center
R_inside = 270 - int(patch_h * np.sqrt(2.0) / 2.0) #radius is 270 (from DRIVE db docs), minus the patch diagonal (assumed it is a square #this is the limit to contain the full patch in the FOV
radius = np.sqrt((x_*x_)+(y_*y_))
if radius < R_inside:
return True
else:
return False
#Divide all the full_imgs in pacthes
def extract_ordered(full_imgs, patch_h, patch_w):
assert (len(full_imgs.shape)==4) #4D arrays
assert (full_imgs.shape[1]==1 or full_imgs.shape[1]==3) #check the channel is 1 or 3
img_h = full_imgs.shape[2] #height of the full image
img_w = full_imgs.shape[3] #width of the full image
N_patches_h = int(img_h/patch_h) #round to lowest int
if (img_h%patch_h != 0):
print "warning: " +str(N_patches_h) +" patches in height, with about " +str(img_h%patch_h) +" pixels left over"
N_patches_w = int(img_w/patch_w) #round to lowest int
if (img_h%patch_h != 0):
print "warning: " +str(N_patches_w) +" patches in width, with about " +str(img_w%patch_w) +" pixels left over"
print "number of patches per image: " +str(N_patches_h*N_patches_w)
N_patches_tot = (N_patches_h*N_patches_w)*full_imgs.shape[0]
patches = np.empty((N_patches_tot,full_imgs.shape[1],patch_h,patch_w))
iter_tot = 0 #iter over the total number of patches (N_patches)
for i in range(full_imgs.shape[0]): #loop over the full images
for h in range(N_patches_h):
for w in range(N_patches_w):
patch = full_imgs[i,:,h*patch_h:(h*patch_h)+patch_h,w*patch_w:(w*patch_w)+patch_w]
patches[iter_tot]=patch
iter_tot +=1 #total
assert (iter_tot==N_patches_tot)
return patches #array with all the full_imgs divided in patches
def paint_border_overlap(full_imgs, patch_h, patch_w, stride_h, stride_w):
assert (len(full_imgs.shape)==4) #4D arrays
assert (full_imgs.shape[1]==1 or full_imgs.shape[1]==3) #check the channel is 1 or 3
img_h = full_imgs.shape[2] #height of the full image
img_w = full_imgs.shape[3] #width of the full image
leftover_h = (img_h-patch_h)%stride_h #leftover on the h dim
leftover_w = (img_w-patch_w)%stride_w #leftover on the w dim
if (leftover_h != 0): #change dimension of img_h
print "\nthe side H is not compatible with the selected stride of " +str(stride_h)
print "img_h " +str(img_h) + ", patch_h " +str(patch_h) + ", stride_h " +str(stride_h)
print "(img_h - patch_h) MOD stride_h: " +str(leftover_h)
print "So the H dim will be padded with additional " +str(stride_h - leftover_h) + " pixels"
tmp_full_imgs = np.zeros((full_imgs.shape[0],full_imgs.shape[1],img_h+(stride_h-leftover_h),img_w))
tmp_full_imgs[0:full_imgs.shape[0],0:full_imgs.shape[1],0:img_h,0:img_w] = full_imgs
full_imgs = tmp_full_imgs
if (leftover_w != 0): #change dimension of img_w
print "the side W is not compatible with the selected stride of " +str(stride_w)
print "img_w " +str(img_w) + ", patch_w " +str(patch_w) + ", stride_w " +str(stride_w)
print "(img_w - patch_w) MOD stride_w: " +str(leftover_w)
print "So the W dim will be padded with additional " +str(stride_w - leftover_w) + " pixels"
tmp_full_imgs = np.zeros((full_imgs.shape[0],full_imgs.shape[1],full_imgs.shape[2],img_w+(stride_w - leftover_w)))
tmp_full_imgs[0:full_imgs.shape[0],0:full_imgs.shape[1],0:full_imgs.shape[2],0:img_w] = full_imgs
full_imgs = tmp_full_imgs
print "new full images shape: \n" +str(full_imgs.shape)
return full_imgs
#Divide all the full_imgs in pacthes
def extract_ordered_overlap(full_imgs, patch_h, patch_w,stride_h,stride_w):
assert (len(full_imgs.shape)==4) #4D arrays
assert (full_imgs.shape[1]==1 or full_imgs.shape[1]==3) #check the channel is 1 or 3
img_h = full_imgs.shape[2] #height of the full image
img_w = full_imgs.shape[3] #width of the full image
assert ((img_h-patch_h)%stride_h==0 and (img_w-patch_w)%stride_w==0)
N_patches_img = ((img_h-patch_h)//stride_h+1)*((img_w-patch_w)//stride_w+1) #// --> division between integers
N_patches_tot = N_patches_img*full_imgs.shape[0]
print "Number of patches on h : " +str(((img_h-patch_h)//stride_h+1))
print "Number of patches on w : " +str(((img_w-patch_w)//stride_w+1))
print "number of patches per image: " +str(N_patches_img) +", totally for this dataset: " +str(N_patches_tot)
patches = np.empty((N_patches_tot,full_imgs.shape[1],patch_h,patch_w))
iter_tot = 0 #iter over the total number of patches (N_patches)
for i in range(full_imgs.shape[0]): #loop over the full images
for h in range((img_h-patch_h)//stride_h+1):
for w in range((img_w-patch_w)//stride_w+1):
patch = full_imgs[i,:,h*stride_h:(h*stride_h)+patch_h,w*stride_w:(w*stride_w)+patch_w]
patches[iter_tot]=patch
iter_tot +=1 #total
assert (iter_tot==N_patches_tot)
return patches #array with all the full_imgs divided in patches
def recompone_overlap(preds, img_h, img_w, stride_h, stride_w):
assert (len(preds.shape)==4) #4D arrays
assert (preds.shape[1]==1 or preds.shape[1]==3) #check the channel is 1 or 3
patch_h = preds.shape[2]
patch_w = preds.shape[3]
N_patches_h = (img_h-patch_h)//stride_h+1
N_patches_w = (img_w-patch_w)//stride_w+1
N_patches_img = N_patches_h * N_patches_w
print "N_patches_h: " +str(N_patches_h)
print "N_patches_w: " +str(N_patches_w)
print "N_patches_img: " +str(N_patches_img)
assert (preds.shape[0]%N_patches_img==0)
N_full_imgs = preds.shape[0]//N_patches_img
print "According to the dimension inserted, there are " +str(N_full_imgs) +" full images (of " +str(img_h)+"x" +str(img_w) +" each)"
full_prob = np.zeros((N_full_imgs,preds.shape[1],img_h,img_w)) #itialize to zero mega array with sum of Probabilities
full_sum = np.zeros((N_full_imgs,preds.shape[1],img_h,img_w))
k = 0 #iterator over all the patches
for i in range(N_full_imgs):
for h in range((img_h-patch_h)//stride_h+1):
for w in range((img_w-patch_w)//stride_w+1):
full_prob[i,:,h*stride_h:(h*stride_h)+patch_h,w*stride_w:(w*stride_w)+patch_w]+=preds[k]
full_sum[i,:,h*stride_h:(h*stride_h)+patch_h,w*stride_w:(w*stride_w)+patch_w]+=1
k+=1
assert(k==preds.shape[0])
assert(np.min(full_sum)>=1.0) #at least one
final_avg = full_prob/full_sum
print final_avg.shape
assert(np.max(final_avg)<=1.0) #max value for a pixel is 1.0
assert(np.min(final_avg)>=0.0) #min value for a pixel is 0.0
return final_avg
#Recompone the full images with the patches
def recompone(data,N_h,N_w):
assert (data.shape[1]==1 or data.shape[1]==3) #check the channel is 1 or 3
assert(len(data.shape)==4)
N_pacth_per_img = N_w*N_h
assert(data.shape[0]%N_pacth_per_img == 0)
N_full_imgs = data.shape[0]/N_pacth_per_img
patch_h = data.shape[2]
patch_w = data.shape[3]
N_pacth_per_img = N_w*N_h
#define and start full recompone
full_recomp = np.empty((N_full_imgs,data.shape[1],N_h*patch_h,N_w*patch_w))
k = 0 #iter full img
s = 0 #iter single patch
while (s<data.shape[0]):
#recompone one:
single_recon = np.empty((data.shape[1],N_h*patch_h,N_w*patch_w))
for h in range(N_h):
for w in range(N_w):
single_recon[:,h*patch_h:(h*patch_h)+patch_h,w*patch_w:(w*patch_w)+patch_w]=data[s]
s+=1
full_recomp[k]=single_recon
k+=1
assert (k==N_full_imgs)
return full_recomp
#Extend the full images becasue patch divison is not exact
def paint_border(data,patch_h,patch_w):
assert (len(data.shape)==4) #4D arrays
assert (data.shape[1]==1 or data.shape[1]==3) #check the channel is 1 or 3
img_h=data.shape[2]
img_w=data.shape[3]
new_img_h = 0
new_img_w = 0
if (img_h%patch_h)==0:
new_img_h = img_h
else:
new_img_h = ((int(img_h)/int(patch_h))+1)*patch_h
if (img_w%patch_w)==0:
new_img_w = img_w
else:
new_img_w = ((int(img_w)/int(patch_w))+1)*patch_w
new_data = np.zeros((data.shape[0],data.shape[1],new_img_h,new_img_w))
new_data[:,:,0:img_h,0:img_w] = data[:,:,:,:]
return new_data
#return only the pixels contained in the FOV, for both images and masks
def pred_only_FOV(data_imgs,data_masks,original_imgs_border_masks):
assert (len(data_imgs.shape)==4 and len(data_masks.shape)==4) #4D arrays
assert (data_imgs.shape[0]==data_masks.shape[0])
assert (data_imgs.shape[2]==data_masks.shape[2])
assert (data_imgs.shape[3]==data_masks.shape[3])
assert (data_imgs.shape[1]==1 and data_masks.shape[1]==1) #check the channel is 1
height = data_imgs.shape[2]
width = data_imgs.shape[3]
new_pred_imgs = []
new_pred_masks = []
for i in range(data_imgs.shape[0]): #loop over the full images
for x in range(width):
for y in range(height):
if inside_FOV_DRIVE(i,x,y,original_imgs_border_masks)==True:
new_pred_imgs.append(data_imgs[i,:,y,x])
new_pred_masks.append(data_masks[i,:,y,x])
new_pred_imgs = np.asarray(new_pred_imgs)
new_pred_masks = np.asarray(new_pred_masks)
return new_pred_imgs, new_pred_masks
#function to set to black everything outside the FOV, in a full image
def kill_border(data, original_imgs_border_masks):
assert (len(data.shape)==4) #4D arrays
assert (data.shape[1]==1 or data.shape[1]==3) #check the channel is 1 or 3
height = data.shape[2]
width = data.shape[3]
for i in range(data.shape[0]): #loop over the full images
for x in range(width):
for y in range(height):
if inside_FOV_DRIVE(i,x,y,original_imgs_border_masks)==False:
data[i,:,y,x]=0.0
def inside_FOV_DRIVE(i, x, y, DRIVE_masks):
assert (len(DRIVE_masks.shape)==4) #4D arrays
assert (DRIVE_masks.shape[1]==1) #DRIVE masks is black and white
# DRIVE_masks = DRIVE_masks/255. #NOOO!! otherwise with float numbers takes forever!!
if (x >= DRIVE_masks.shape[3] or y >= DRIVE_masks.shape[2]): #my image bigger than the original
return False
if (DRIVE_masks[i,0,y,x]>0): #0==black pixels
# print DRIVE_masks[i,0,y,x] #verify it is working right
return True
else:
return False
================================================
FILE: lib/help_functions.py
================================================
import h5py
import numpy as np
from PIL import Image
from matplotlib import pyplot as plt
def load_hdf5(infile):
with h5py.File(infile,"r") as f: #"with" close the file after its nested commands
return f["image"][()]
def write_hdf5(arr,outfile):
with h5py.File(outfile,"w") as f:
f.create_dataset("image", data=arr, dtype=arr.dtype)
#convert RGB image in black and white
def rgb2gray(rgb):
assert (len(rgb.shape)==4) #4D arrays
assert (rgb.shape[1]==3)
bn_imgs = rgb[:,0,:,:]*0.299 + rgb[:,1,:,:]*0.587 + rgb[:,2,:,:]*0.114
bn_imgs = np.reshape(bn_imgs,(rgb.shape[0],1,rgb.shape[2],rgb.shape[3]))
return bn_imgs
#group a set of images row per columns
def group_images(data,per_row):
assert data.shape[0]%per_row==0
assert (data.shape[1]==1 or data.shape[1]==3)
data = np.transpose(data,(0,2,3,1)) #corect format for imshow
all_stripe = []
for i in range(int(data.shape[0]/per_row)):
stripe = data[i*per_row]
for k in range(i*per_row+1, i*per_row+per_row):
stripe = np.concatenate((stripe,data[k]),axis=1)
all_stripe.append(stripe)
totimg = all_stripe[0]
for i in range(1,len(all_stripe)):
totimg = np.concatenate((totimg,all_stripe[i]),axis=0)
return totimg
#visualize image (as PIL image, NOT as matplotlib!)
def visualize(data,filename):
assert (len(data.shape)==3) #height*width*channels
img = None
if data.shape[2]==1: #in case it is black and white
data = np.reshape(data,(data.shape[0],data.shape[1]))
if np.max(data)>1:
img = Image.fromarray(data.astype(np.uint8)) #the image is already 0-255
else:
img = Image.fromarray((data*255).astype(np.uint8)) #the image is between 0-1
img.save(filename + '.png')
return img
#prepare the mask in the right shape for the Unet
def masks_Unet(masks):
assert (len(masks.shape)==4) #4D arrays
assert (masks.shape[1]==1 ) #check the channel is 1
im_h = masks.shape[2]
im_w = masks.shape[3]
masks = np.reshape(masks,(masks.shape[0],im_h*im_w))
new_masks = np.empty((masks.shape[0],im_h*im_w,2))
#new_masks[np.where(masks==0),0]=1
#new_masks[np.where(masks==1),1]=1
for i in range(masks.shape[0]):
for j in range(im_h*im_w):
if masks[i,j] == 0:
new_masks[i,j,0]=1
new_masks[i,j,1]=0
else:
new_masks[i,j,0]=0
new_masks[i,j,1]=1
return new_masks
def pred_to_imgs(pred, patch_height, patch_width, mode="original"):
assert (len(pred.shape)==3) #3D array: (Npatches,height*width,2)
assert (pred.shape[2]==2 ) #check the classes are 2
pred_images = np.empty((pred.shape[0],pred.shape[1])) #(Npatches,height*width)
if mode=="original":
for i in range(pred.shape[0]):
for pix in range(pred.shape[1]):
pred_images[i,pix]=pred[i,pix,1]
elif mode=="threshold":
for i in range(pred.shape[0]):
for pix in range(pred.shape[1]):
if pred[i,pix,1]>=0.5:
pred_images[i,pix]=1
else:
pred_images[i,pix]=0
else:
print("mode " +str(mode) +" not recognized, it can be 'original' or 'threshold'")
exit()
pred_images = np.reshape(pred_images,(pred_images.shape[0],1, patch_height, patch_width))
return pred_images
================================================
FILE: lib/help_functions.py.bak
================================================
import h5py
import numpy as np
from PIL import Image
from matplotlib import pyplot as plt
def load_hdf5(infile):
with h5py.File(infile,"r") as f: #"with" close the file after its nested commands
return f["image"][()]
def write_hdf5(arr,outfile):
with h5py.File(outfile,"w") as f:
f.create_dataset("image", data=arr, dtype=arr.dtype)
#convert RGB image in black and white
def rgb2gray(rgb):
assert (len(rgb.shape)==4) #4D arrays
assert (rgb.shape[1]==3)
bn_imgs = rgb[:,0,:,:]*0.299 + rgb[:,1,:,:]*0.587 + rgb[:,2,:,:]*0.114
bn_imgs = np.reshape(bn_imgs,(rgb.shape[0],1,rgb.shape[2],rgb.shape[3]))
return bn_imgs
#group a set of images row per columns
def group_images(data,per_row):
assert data.shape[0]%per_row==0
assert (data.shape[1]==1 or data.shape[1]==3)
data = np.transpose(data,(0,2,3,1)) #corect format for imshow
all_stripe = []
for i in range(int(data.shape[0]/per_row)):
stripe = data[i*per_row]
for k in range(i*per_row+1, i*per_row+per_row):
stripe = np.concatenate((stripe,data[k]),axis=1)
all_stripe.append(stripe)
totimg = all_stripe[0]
for i in range(1,len(all_stripe)):
totimg = np.concatenate((totimg,all_stripe[i]),axis=0)
return totimg
#visualize image (as PIL image, NOT as matplotlib!)
def visualize(data,filename):
assert (len(data.shape)==3) #height*width*channels
img = None
if data.shape[2]==1: #in case it is black and white
data = np.reshape(data,(data.shape[0],data.shape[1]))
if np.max(data)>1:
img = Image.fromarray(data.astype(np.uint8)) #the image is already 0-255
else:
img = Image.fromarray((data*255).astype(np.uint8)) #the image is between 0-1
img.save(filename + '.png')
return img
#prepare the mask in the right shape for the Unet
def masks_Unet(masks):
assert (len(masks.shape)==4) #4D arrays
assert (masks.shape[1]==1 ) #check the channel is 1
im_h = masks.shape[2]
im_w = masks.shape[3]
masks = np.reshape(masks,(masks.shape[0],im_h*im_w))
new_masks = np.empty((masks.shape[0],im_h*im_w,2))
for i in range(masks.shape[0]):
for j in range(im_h*im_w):
if masks[i,j] == 0:
new_masks[i,j,0]=1
new_masks[i,j,1]=0
else:
new_masks[i,j,0]=0
new_masks[i,j,1]=1
return new_masks
def pred_to_imgs(pred, patch_height, patch_width, mode="original"):
assert (len(pred.shape)==3) #3D array: (Npatches,height*width,2)
assert (pred.shape[2]==2 ) #check the classes are 2
pred_images = np.empty((pred.shape[0],pred.shape[1])) #(Npatches,height*width)
if mode=="original":
for i in range(pred.shape[0]):
for pix in range(pred.shape[1]):
pred_images[i,pix]=pred[i,pix,1]
elif mode=="threshold":
for i in range(pred.shape[0]):
for pix in range(pred.shape[1]):
if pred[i,pix,1]>=0.5:
pred_images[i,pix]=1
else:
pred_images[i,pix]=0
else:
print "mode " +str(mode) +" not recognized, it can be 'original' or 'threshold'"
exit()
pred_images = np.reshape(pred_images,(pred_images.shape[0],1, patch_height, patch_width))
return pred_images
================================================
FILE: lib/pre_processing.py
================================================
###################################################
#
# Script to pre-process the original imgs
#
##################################################
import numpy as np
from PIL import Image
import cv2
from .help_functions import *
#My pre processing (use for both training and testing!)
def my_PreProc(data):
assert(len(data.shape)==4)
assert (data.shape[1]==3) #Use the original images
#black-white conversion
train_imgs = rgb2gray(data)
#my preprocessing:
train_imgs = dataset_normalized(train_imgs)
train_imgs = clahe_equalized(train_imgs)
train_imgs = adjust_gamma(train_imgs, 1.2)
train_imgs = train_imgs/255. #reduce to 0-1 range
return train_imgs
#============================================================
#========= PRE PROCESSING FUNCTIONS ========================#
#============================================================
#==== histogram equalization
def histo_equalized(imgs):
assert (len(imgs.shape)==4) #4D arrays
assert (imgs.shape[1]==1) #check the channel is 1
imgs_equalized = np.empty(imgs.shape)
for i in range(imgs.shape[0]):
imgs_equalized[i,0] = cv2.equalizeHist(np.array(imgs[i,0], dtype = np.uint8))
return imgs_equalized
# CLAHE (Contrast Limited Adaptive Histogram Equalization)
#adaptive histogram equalization is used. In this, image is divided into small blocks called "tiles" (tileSize is 8x8 by default in OpenCV). Then each of these blocks are histogram equalized as usual. So in a small area, histogram would confine to a small region (unless there is noise). If noise is there, it will be amplified. To avoid this, contrast limiting is applied. If any histogram bin is above the specified contrast limit (by default 40 in OpenCV), those pixels are clipped and distributed uniformly to other bins before applying histogram equalization. After equalization, to remove artifacts in tile borders, bilinear interpolation is applied
def clahe_equalized(imgs):
assert (len(imgs.shape)==4) #4D arrays
assert (imgs.shape[1]==1) #check the channel is 1
#create a CLAHE object (Arguments are optional).
clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8,8))
imgs_equalized = np.empty(imgs.shape)
for i in range(imgs.shape[0]):
imgs_equalized[i,0] = clahe.apply(np.array(imgs[i,0], dtype = np.uint8))
return imgs_equalized
# ===== normalize over the dataset
def dataset_normalized(imgs):
assert (len(imgs.shape)==4) #4D arrays
assert (imgs.shape[1]==1) #check the channel is 1
imgs_normalized = np.empty(imgs.shape)
imgs_std = np.std(imgs)
imgs_mean = np.mean(imgs)
imgs_normalized = (imgs-imgs_mean)/imgs_std
for i in range(imgs.shape[0]):
imgs_normalized[i] = ((imgs_normalized[i] - np.min(imgs_normalized[i])) / (np.max(imgs_normalized[i])-np.min(imgs_normalized[i])))*255
return imgs_normalized
def adjust_gamma(imgs, gamma=1.0):
assert (len(imgs.shape)==4) #4D arrays
assert (imgs.shape[1]==1) #check the channel is 1
# build a lookup table mapping the pixel values [0, 255] to
# their adjusted gamma values
invGamma = 1.0 / gamma
table = np.array([((i / 255.0) ** invGamma) * 255 for i in np.arange(0, 256)]).astype("uint8")
# apply gamma correction using the lookup table
new_imgs = np.empty(imgs.shape)
for i in range(imgs.shape[0]):
new_imgs[i,0] = cv2.LUT(np.array(imgs[i,0], dtype = np.uint8), table)
return new_imgs
================================================
FILE: lib/pre_processing.py.bak
================================================
###################################################
#
# Script to pre-process the original imgs
#
##################################################
import numpy as np
from PIL import Image
import cv2
from help_functions import *
#My pre processing (use for both training and testing!)
def my_PreProc(data):
assert(len(data.shape)==4)
assert (data.shape[1]==3) #Use the original images
#black-white conversion
train_imgs = rgb2gray(data)
#my preprocessing:
train_imgs = dataset_normalized(train_imgs)
train_imgs = clahe_equalized(train_imgs)
train_imgs = adjust_gamma(train_imgs, 1.2)
train_imgs = train_imgs/255. #reduce to 0-1 range
return train_imgs
#============================================================
#========= PRE PROCESSING FUNCTIONS ========================#
#============================================================
#==== histogram equalization
def histo_equalized(imgs):
assert (len(imgs.shape)==4) #4D arrays
assert (imgs.shape[1]==1) #check the channel is 1
imgs_equalized = np.empty(imgs.shape)
for i in range(imgs.shape[0]):
imgs_equalized[i,0] = cv2.equalizeHist(np.array(imgs[i,0], dtype = np.uint8))
return imgs_equalized
# CLAHE (Contrast Limited Adaptive Histogram Equalization)
#adaptive histogram equalization is used. In this, image is divided into small blocks called "tiles" (tileSize is 8x8 by default in OpenCV). Then each of these blocks are histogram equalized as usual. So in a small area, histogram would confine to a small region (unless there is noise). If noise is there, it will be amplified. To avoid this, contrast limiting is applied. If any histogram bin is above the specified contrast limit (by default 40 in OpenCV), those pixels are clipped and distributed uniformly to other bins before applying histogram equalization. After equalization, to remove artifacts in tile borders, bilinear interpolation is applied
def clahe_equalized(imgs):
assert (len(imgs.shape)==4) #4D arrays
assert (imgs.shape[1]==1) #check the channel is 1
#create a CLAHE object (Arguments are optional).
clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8,8))
imgs_equalized = np.empty(imgs.shape)
for i in range(imgs.shape[0]):
imgs_equalized[i,0] = clahe.apply(np.array(imgs[i,0], dtype = np.uint8))
return imgs_equalized
# ===== normalize over the dataset
def dataset_normalized(imgs):
assert (len(imgs.shape)==4) #4D arrays
assert (imgs.shape[1]==1) #check the channel is 1
imgs_normalized = np.empty(imgs.shape)
imgs_std = np.std(imgs)
imgs_mean = np.mean(imgs)
imgs_normalized = (imgs-imgs_mean)/imgs_std
for i in range(imgs.shape[0]):
imgs_normalized[i] = ((imgs_normalized[i] - np.min(imgs_normalized[i])) / (np.max(imgs_normalized[i])-np.min(imgs_normalized[i])))*255
return imgs_normalized
def adjust_gamma(imgs, gamma=1.0):
assert (len(imgs.shape)==4) #4D arrays
assert (imgs.shape[1]==1) #check the channel is 1
# build a lookup table mapping the pixel values [0, 255] to
# their adjusted gamma values
invGamma = 1.0 / gamma
table = np.array([((i / 255.0) ** invGamma) * 255 for i in np.arange(0, 256)]).astype("uint8")
# apply gamma correction using the lookup table
new_imgs = np.empty(imgs.shape)
for i in range(imgs.shape[0]):
new_imgs[i,0] = cv2.LUT(np.array(imgs[i,0], dtype = np.uint8), table)
return new_imgs
================================================
FILE: prepare_datasets_DRIVE.py
================================================
#==========================================================
#
# This prepare the hdf5 datasets of the DRIVE database
#
#============================================================
import os
import h5py
import numpy as np
from PIL import Image
def write_hdf5(arr,outfile):
with h5py.File(outfile,"w") as f:
f.create_dataset("image", data=arr, dtype=arr.dtype)
#------------Path of the images --------------------------------------------------------------
#train
original_imgs_train = "./DRIVE/training/images/"
groundTruth_imgs_train = "./DRIVE/training/1st_manual/"
borderMasks_imgs_train = "./DRIVE/training/mask/"
#test
original_imgs_test = "./DRIVE/test/images/"
groundTruth_imgs_test = "./DRIVE/test/1st_manual/"
borderMasks_imgs_test = "./DRIVE/test/mask/"
#---------------------------------------------------------------------------------------------
Nimgs = 20
channels = 3
height = 584
width = 565
dataset_path = "./DRIVE_datasets_training_testing/"
if not os.path.isdir(dataset_path):
os.mkdir(dataset_path)
def get_datasets(imgs_dir,groundTruth_dir,borderMasks_dir,train_test="null"):
imgs = np.empty((Nimgs,height,width,channels))
groundTruth = np.empty((Nimgs,height,width))
border_masks = np.empty((Nimgs,height,width))
for path, subdirs, files in os.walk(imgs_dir): #list all files, directories in the path
for i in range(len(files)):
#original
print("original image: " +files[i])
img = Image.open(imgs_dir+files[i])
imgs[i] = np.asarray(img)
#corresponding ground truth
groundTruth_name = files[i][0:2] + "_manual1.gif"
print("ground truth name: " + groundTruth_name)
g_truth = Image.open(groundTruth_dir + groundTruth_name)
groundTruth[i] = np.asarray(g_truth)
#corresponding border masks
border_masks_name = ""
if train_test=="train":
border_masks_name = files[i][0:2] + "_training_mask.gif"
elif train_test=="test":
border_masks_name = files[i][0:2] + "_test_mask.gif"
else:
print("specify if train or test!!")
exit()
print("border masks name: " + border_masks_name)
b_mask = Image.open(borderMasks_dir + border_masks_name)
border_masks[i] = np.asarray(b_mask)
print("imgs max: " +str(np.max(imgs)))
print("imgs min: " +str(np.min(imgs)))
assert(np.max(groundTruth)==255 and np.max(border_masks)==255)
assert(np.min(groundTruth)==0 and np.min(border_masks)==0)
print("ground truth and border masks are correctly withih pixel value range 0-255 (black-white)")
#reshaping for my standard tensors
imgs = np.transpose(imgs,(0,3,1,2))
assert(imgs.shape == (Nimgs,channels,height,width))
groundTruth = np.reshape(groundTruth,(Nimgs,1,height,width))
border_masks = np.reshape(border_masks,(Nimgs,1,height,width))
assert(groundTruth.shape == (Nimgs,1,height,width))
assert(border_masks.shape == (Nimgs,1,height,width))
return imgs, groundTruth, border_masks
#getting the training datasets
imgs_train, groundTruth_train, border_masks_train = get_datasets(original_imgs_train,groundTruth_imgs_train,borderMasks_imgs_train,"train")
print("saving train datasets")
write_hdf5(imgs_train, dataset_path + "DRIVE_dataset_imgs_train.hdf5")
write_hdf5(groundTruth_train, dataset_path + "DRIVE_dataset_groundTruth_train.hdf5")
write_hdf5(border_masks_train,dataset_path + "DRIVE_dataset_borderMasks_train.hdf5")
#getting the testing datasets
imgs_test, groundTruth_test, border_masks_test = get_datasets(original_imgs_test,groundTruth_imgs_test,borderMasks_imgs_test,"test")
print("saving test datasets")
write_hdf5(imgs_test,dataset_path + "DRIVE_dataset_imgs_test.hdf5")
write_hdf5(groundTruth_test, dataset_path + "DRIVE_dataset_groundTruth_test.hdf5")
write_hdf5(border_masks_test,dataset_path + "DRIVE_dataset_borderMasks_test.hdf5")
================================================
FILE: src/LadderNetv65.py
================================================
import torch
import torch.nn.functional as F
import torch.nn as nn
drop = 0.25
def conv3x3(in_planes, out_planes, stride=1):
"""3x3 convolution with padding"""
return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,
padding=1, bias=True)
class BasicBlock(nn.Module):
expansion = 1
def __init__(self, inplanes, planes, stride=1, downsample=None):
super(BasicBlock, self).__init__()
if inplanes!= planes:
self.conv0 = conv3x3(inplanes,planes)
self.inplanes = inplanes
self.planes = planes
self.conv1 = conv3x3(planes, planes, stride)
#self.bn1 = nn.BatchNorm2d(planes)
self.relu = nn.ReLU(inplace=True)
#self.conv2 = conv3x3(planes, planes)
#self.bn2 = nn.BatchNorm2d(planes)
self.downsample = downsample
self.stride = stride
self.drop = nn.Dropout2d(p=drop)
def forward(self, x):
if self.inplanes != self.planes:
x = self.conv0(x)
x = F.relu(x)
out = self.conv1(x)
#out = self.bn1(out)
out = self.relu(out)
out = self.drop(out)
out1 = self.conv1(out)
#out1 = self.relu(out1)
out2 = out1 + x
return F.relu(out2)
class Bottleneck(nn.Module):
expansion = 4
def __init__(self, inplanes, planes, stride=1, downsample=None):
super(Bottleneck, self).__init__()
self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1, bias=False)
self.bn1 = nn.BatchNorm2d(planes)
self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride,
padding=1, bias=False)
self.bn2 = nn.BatchNorm2d(planes)
self.conv3 = nn.Conv2d(planes, planes * self.expansion, kernel_size=1, bias=False)
self.bn3 = nn.BatchNorm2d(planes * self.expansion)
self.relu = nn.ReLU(inplace=True)
self.downsample = downsample
self.stride = stride
def forward(self, x):
residual = x
out = self.conv1(x)
out = self.bn1(out)
out = self.relu(out)
out = self.conv2(out)
out = self.bn2(out)
out = self.relu(out)
out = self.conv3(out)
out = self.bn3(out)
if self.downsample is not None:
residual = self.downsample(x)
out += residual
out = self.relu(out)
return out
class Initial_LadderBlock(nn.Module):
def __init__(self,planes,layers,kernel=3,block=BasicBlock,inplanes = 3):
super().__init__()
self.planes = planes
self.layers = layers
self.kernel = kernel
self.padding = int((kernel-1)/2)
self.inconv = nn.Conv2d(in_channels=inplanes,out_channels=planes,
kernel_size=3,stride=1,padding=1,bias=True)
# create module list for down branch
self.down_module_list = nn.ModuleList()
for i in range(0,layers):
self.down_module_list.append(block(planes*(2**i),planes*(2**i)))
# use strided conv instead of poooling
self.down_conv_list = nn.ModuleList()
for i in range(0,layers):
self.down_conv_list.append(nn.Conv2d(planes*2**i,planes*2**(i+1),stride=2,kernel_size=kernel,padding=self.padding))
# create module for bottom block
self.bottom = block(planes*(2**layers),planes*(2**layers))
# create module list for up branch
self.up_conv_list = nn.ModuleList()
self.up_dense_list = nn.ModuleList()
for i in range(0, layers):
self.up_conv_list.append(nn.ConvTranspose2d(in_channels=planes*2**(layers-i), out_channels=planes*2**max(0,layers-i-1), kernel_size=3,
stride=2,padding=1,output_padding=1,bias=True))
self.up_dense_list.append(block(planes*2**max(0,layers-i-1),planes*2**max(0,layers-i-1)))
def forward(self, x):
out = self.inconv(x)
out = F.relu(out)
down_out = []
# down branch
for i in range(0,self.layers):
out = self.down_module_list[i](out)
down_out.append(out)
out = self.down_conv_list[i](out)
out = F.relu(out)
# bottom branch
out = self.bottom(out)
bottom = out
# up branch
up_out = []
up_out.append(bottom)
for j in range(0,self.layers):
out = self.up_conv_list[j](out) + down_out[self.layers-j-1]
#out = F.relu(out)
out = self.up_dense_list[j](out)
up_out.append(out)
return up_out
class LadderBlock(nn.Module):
def __init__(self,planes,layers,kernel=3,block=BasicBlock,inplanes = 3):
super().__init__()
self.planes = planes
self.layers = layers
self.kernel = kernel
self.padding = int((kernel-1)/2)
self.inconv = block(planes,planes)
# create module list for down branch
self.down_module_list = nn.ModuleList()
for i in range(0,layers):
self.down_module_list.append(block(planes*(2**i),planes*(2**i)))
# use strided conv instead of poooling
self.down_conv_list = nn.ModuleList()
for i in range(0,layers):
self.down_conv_list.append(nn.Conv2d(planes*2**i,planes*2**(i+1),stride=2,kernel_size=kernel,padding=self.padding))
# create module for bottom block
self.bottom = block(planes*(2**layers),planes*(2**layers))
# create module list for up branch
self.up_conv_list = nn.ModuleList()
self.up_dense_list = nn.ModuleList()
for i in range(0, layers):
self.up_conv_list.append(nn.ConvTranspose2d(planes*2**(layers-i), planes*2**max(0,layers-i-1), kernel_size=3,
stride=2,padding=1,output_padding=1,bias=True))
self.up_dense_list.append(block(planes*2**max(0,layers-i-1),planes*2**max(0,layers-i-1)))
def forward(self, x):
out = self.inconv(x[-1])
down_out = []
# down branch
for i in range(0,self.layers):
out = out + x[-i-1]
out = self.down_module_list[i](out)
down_out.append(out)
out = self.down_conv_list[i](out)
out = F.relu(out)
# bottom branch
out = self.bottom(out)
bottom = out
# up branch
up_out = []
up_out.append(bottom)
for j in range(0,self.layers):
out = self.up_conv_list[j](out) + down_out[self.layers-j-1]
#out = F.relu(out)
out = self.up_dense_list[j](out)
up_out.append(out)
return up_out
class Final_LadderBlock(nn.Module):
def __init__(self,planes,layers,kernel=3,block=BasicBlock,inplanes = 3):
super().__init__()
self.block = LadderBlock(planes,layers,kernel=kernel,block=block)
def forward(self, x):
out = self.block(x)
return out[-1]
class LadderNetv6(nn.Module):
def __init__(self,layers=3,filters=16,num_classes=2,inplanes=3):
super().__init__()
self.initial_block = Initial_LadderBlock(planes=filters,layers=layers,inplanes=inplanes)
#self.middle_block = LadderBlock(planes=filters,layers=layers)
self.final_block = Final_LadderBlock(planes=filters,layers=layers)
self.final = nn.Conv2d(in_channels=filters,out_channels=num_classes,kernel_size=1)
def forward(self,x):
out = self.initial_block(x)
#out = self.middle_block(out)
out = self.final_block(out)
out = self.final(out)
#out = F.relu(out)
out = F.log_softmax(out,dim=1)
return out
================================================
FILE: src/__init__.py
================================================
================================================
FILE: src/losses.py
================================================
import torch
import numpy as np
import torch.nn as nn
def cuda(x):
return x.cuda(async=True) if torch.cuda.is_available() else x
class LossMulti:
def __init__(self, jaccard_weight=0, class_weights=None, num_classes=1):
if class_weights is not None:
nll_weight = cuda(
torch.from_numpy(class_weights.astype(np.float32)))
else:
nll_weight = None
self.nll_loss = nn.NLLLoss2d(weight=nll_weight)
self.jaccard_weight = jaccard_weight
self.num_classes = num_classes
def __call__(self, outputs, targets):
loss = (1 - self.jaccard_weight) * self.nll_loss(outputs, targets)
if self.jaccard_weight:
eps = 1e-15
for cls in range(self.num_classes):
jaccard_target = (targets == cls).float()
jaccard_output = outputs[:, cls].exp()
intersection = (jaccard_output * jaccard_target).sum()
union = jaccard_output.sum() + jaccard_target.sum()
loss -= torch.log((intersection + eps) / (union - intersection + eps)) * self.jaccard_weight
return loss
================================================
FILE: src/retinaNN_predict.py
================================================
###################################################
#
# Script to
# - Calculate prediction of the test dataset
# - Calculate the parameters to evaluate the prediction
#
##################################################
#Python
import numpy as np
import configparser
from matplotlib import pyplot as plt
import torch
import os
import torch.backends.cudnn as cudnn
from torch.utils.data import DataLoader,Dataset
#scikit learn
from sklearn.metrics import roc_curve
from sklearn.metrics import roc_auc_score
from sklearn.metrics import confusion_matrix
from sklearn.metrics import precision_recall_curve
from sklearn.metrics import jaccard_similarity_score
from sklearn.metrics import f1_score
import sys
sys.path.insert(0, '../')
# help_functions.py
from lib.help_functions import *
# extract_patches.py
from lib.extract_patches import recompone
from lib.extract_patches import recompone_overlap
from lib.extract_patches import paint_border
from lib.extract_patches import kill_border
from lib.extract_patches import pred_only_FOV
from lib.extract_patches import get_data_testing
from lib.extract_patches import get_data_testing_overlap
# pre_processing.py
from lib.pre_processing import my_PreProc
# define pyplot parameters
import matplotlib.pylab as pylab
params = {'legend.fontsize': 15,
'axes.labelsize': 15,
'axes.titlesize':15,
'xtick.labelsize':15,
'ytick.labelsize':15}
pylab.rcParams.update(params)
#========= CONFIG FILE TO READ FROM =======
config = configparser.RawConfigParser()
config.read('../configuration.txt')
#===========================================
#run the training on invariant or local
path_data = config.get('data paths', 'path_local')
#original test images (for FOV selection)
DRIVE_test_imgs_original = path_data + config.get('data paths', 'test_imgs_original')
test_imgs_orig = load_hdf5(DRIVE_test_imgs_original)
full_img_height = test_imgs_orig.shape[2]
full_img_width = test_imgs_orig.shape[3]
#the border masks provided by the DRIVE
DRIVE_test_border_masks = path_data + config.get('data paths', 'test_border_masks')
test_border_masks = load_hdf5(DRIVE_test_border_masks)
# dimension of the patches
patch_height = int(config.get('data attributes', 'patch_height'))
patch_width = int(config.get('data attributes', 'patch_width'))
#the stride in case output with average
stride_height = int(config.get('testing settings', 'stride_height'))
stride_width = int(config.get('testing settings', 'stride_width'))
assert (stride_height < patch_height and stride_width < patch_width)
#model name
name_experiment = config.get('experiment name', 'name')
path_experiment = '../' +name_experiment +'/'
#N full images to be predicted
Imgs_to_test = int(config.get('testing settings', 'full_images_to_test'))
#Grouping of the predicted images
N_visual = int(config.get('testing settings', 'N_group_visual'))
#====== average mode ===========
average_mode = config.getboolean('testing settings', 'average_mode')
# #ground truth
# gtruth= path_data + config.get('data paths', 'test_groundTruth')
# img_truth= load_hdf5(gtruth)
# visualize(group_images(test_imgs_orig[0:20,:,:,:],5),'original')#.show()
# visualize(group_images(test_border_masks[0:20,:,:,:],5),'borders')#.show()
# visualize(group_images(img_truth[0:20,:,:,:],5),'gtruth')#.show()
#============ Load the data and divide in patches
patches_imgs_test = None
new_height = None
new_width = None
masks_test = None
patches_masks_test = None
if average_mode == True:
patches_imgs_test, new_height, new_width, masks_test = get_data_testing_overlap(
DRIVE_test_imgs_original = DRIVE_test_imgs_original, #original
DRIVE_test_groudTruth = path_data + config.get('data paths', 'test_groundTruth'), #masks
Imgs_to_test = int(config.get('testing settings', 'full_images_to_test')),
patch_height = patch_height,
patch_width = patch_width,
stride_height = stride_height,
stride_width = stride_width
)
else:
patches_imgs_test, patches_masks_test = get_data_testing(
DRIVE_test_imgs_original = DRIVE_test_imgs_original, #original
DRIVE_test_groudTruth = path_data + config.get('data paths', 'test_groundTruth'), #masks
Imgs_to_test = int(config.get('testing settings', 'full_images_to_test')),
patch_height = patch_height,
patch_width = patch_width,
)
#================ Run the prediction of the patches ==================================
best_last = config.get('testing settings', 'best_last')
layers= 4
filters =10
check_path = 'LadderNetv65_layer_%d_filter_%d.pt7'% (layers,filters) #'UNet16.pt7'#'UNet_Resnet101.pt7'
from LadderNetv65 import *
net = LadderNetv6(num_classes=2,layers=layers,filters=filters,inplanes=1)
device = 'cuda' if torch.cuda.is_available() else 'cpu'
resume = True
if device == 'cuda':
net.cuda()
net = torch.nn.DataParallel(net, device_ids=range(torch.cuda.device_count()))
cudnn.benchmark = True
if resume:
# Load checkpoint.
print('==> Resuming from checkpoint..')
assert os.path.isdir('checkpoint'), 'Error: no checkpoint directory found!'
checkpoint = torch.load('./checkpoint/'+check_path)
net.load_state_dict(checkpoint['net'])
start_epoch = checkpoint['epoch']
class TrainDataset(Dataset):
"""Endovis 2018 dataset."""
def __init__(self, patches_imgs):
self.imgs = patches_imgs
def __len__(self):
return self.imgs.shape[0]
def __getitem__(self, idx):
return torch.from_numpy(self.imgs[idx,...]).float()
batch_size = 1024
test_set = TrainDataset(patches_imgs_test)
test_loader = DataLoader(test_set, batch_size=batch_size,
shuffle=False, num_workers=4)
preds = []
for batch_idx, inputs in enumerate((test_loader)):
inputs = inputs.to(device)
outputs = net(inputs)
outputs = torch.nn.functional.softmax(outputs,dim=1)
outputs = outputs.permute(0,2,3,1)
shape = list(outputs.shape)
outputs = outputs.view(-1,shape[1]*shape[2],2)
outputs = outputs.data.cpu().numpy()
preds.append(outputs)
predictions = np.concatenate(preds,axis=0)
print("Predictions finished")
#===== Convert the prediction arrays in corresponding images
pred_patches = pred_to_imgs(predictions, patch_height, patch_width, "original")
#========== Elaborate and visualize the predicted images ====================
pred_imgs = None
orig_imgs = None
gtruth_masks = None
if average_mode == True:
pred_imgs = recompone_overlap(pred_patches, new_height, new_width, stride_height, stride_width)# predictions
orig_imgs = my_PreProc(test_imgs_orig[0:pred_imgs.shape[0],:,:,:]) #originals
gtruth_masks = masks_test #ground truth masks
else:
pred_imgs = recompone(pred_patches,13,12) # predictions
orig_imgs = recompone(patches_imgs_test,13,12) # originals
gtruth_masks = recompone(patches_masks_test,13,12) #masks
# apply the DRIVE masks on the repdictions #set everything outside the FOV to zero!!
kill_border(pred_imgs, test_border_masks) #DRIVE MASK #only for visualization
## back to original dimensions
orig_imgs = orig_imgs[:,:,0:full_img_height,0:full_img_width]
pred_imgs = pred_imgs[:,:,0:full_img_height,0:full_img_width]
gtruth_masks = gtruth_masks[:,:,0:full_img_height,0:full_img_width]
print("Orig imgs shape: " +str(orig_imgs.shape))
print("pred imgs shape: " +str(pred_imgs.shape))
print("Gtruth imgs shape: " +str(gtruth_masks.shape))
visualize(group_images(orig_imgs,N_visual),path_experiment+"all_originals")#.show()
visualize(group_images(pred_imgs,N_visual),path_experiment+"all_predictions")#.show()
visualize(group_images(gtruth_masks,N_visual),path_experiment+"all_groundTruths")#.show()
#visualize results comparing mask and prediction:
assert (orig_imgs.shape[0]==pred_imgs.shape[0] and orig_imgs.shape[0]==gtruth_masks.shape[0])
N_predicted = orig_imgs.shape[0]
group = N_visual
assert (N_predicted%group==0)
for i in range(int(N_predicted/group)):
orig_stripe = group_images(orig_imgs[i*group:(i*group)+group,:,:,:],group)
masks_stripe = group_images(gtruth_masks[i*group:(i*group)+group,:,:,:],group)
pred_stripe = group_images(pred_imgs[i*group:(i*group)+group,:,:,:],group)
total_img = np.concatenate((orig_stripe,masks_stripe,pred_stripe),axis=0)
visualize(total_img,path_experiment+name_experiment +"_Original_GroundTruth_Prediction"+str(i))#.show()
#====== Evaluate the results
print("\n\n======== Evaluate the results =======================")
#predictions only inside the FOV
y_scores, y_true = pred_only_FOV(pred_imgs,gtruth_masks, test_border_masks) #returns data only inside the FOV
print("Calculating results only inside the FOV:")
print("y scores pixels: " +str(y_scores.shape[0]) +" (radius 270: 270*270*3.14==228906), including background around retina: " +str(pred_imgs.shape[0]*pred_imgs.shape[2]*pred_imgs.shape[3]) +" (584*565==329960)")
print("y true pixels: " +str(y_true.shape[0]) +" (radius 270: 270*270*3.14==228906), including background around retina: " +str(gtruth_masks.shape[2]*gtruth_masks.shape[3]*gtruth_masks.shape[0])+" (584*565==329960)")
#Area under the ROC curve
fpr, tpr, thresholds = roc_curve((y_true), y_scores)
AUC_ROC = roc_auc_score(y_true, y_scores)
# test_integral = np.trapz(tpr,fpr) #trapz is numpy integration
print("\nArea under the ROC curve: " +str(AUC_ROC))
roc_curve =plt.figure()
plt.plot(fpr,tpr,'-',label='Area Under the Curve (AUC = %0.4f)' % AUC_ROC)
plt.title('ROC curve')
plt.xlabel("FPR (False Positive Rate)")
plt.ylabel("TPR (True Positive Rate)")
plt.legend(loc="lower right")
plt.savefig(path_experiment+"ROC.png")
#Precision-recall curve
precision, recall, thresholds = precision_recall_curve(y_true, y_scores)
precision = np.fliplr([precision])[0] #so the array is increasing (you won't get negative AUC)
recall = np.fliplr([recall])[0] #so the array is increasing (you won't get negative AUC)
AUC_prec_rec = np.trapz(precision,recall)
print("\nArea under Precision-Recall curve: " +str(AUC_prec_rec))
prec_rec_curve = plt.figure()
plt.plot(recall,precision,'-',label='Area Under the Curve (AUC = %0.4f)' % AUC_prec_rec)
plt.title('Precision - Recall curve')
plt.xlabel("Recall")
plt.ylabel("Precision")
plt.legend(loc="lower right")
plt.savefig(path_experiment+"Precision_recall.png")
#Confusion matrix
threshold_confusion = 0.5
print("\nConfusion matrix: Costum threshold (for positive) of " +str(threshold_confusion))
y_pred = np.empty((y_scores.shape[0]))
for i in range(y_scores.shape[0]):
if y_scores[i]>=threshold_confusion:
y_pred[i]=1
else:
y_pred[i]=0
confusion = confusion_matrix(y_true, y_pred)
print(confusion)
accuracy = 0
if float(np.sum(confusion))!=0:
accuracy = float(confusion[0,0]+confusion[1,1])/float(np.sum(confusion))
print("Global Accuracy: " +str(accuracy))
specificity = 0
if float(confusion[0,0]+confusion[0,1])!=0:
specificity = float(confusion[0,0])/float(confusion[0,0]+confusion[0,1])
print("Specificity: " +str(specificity))
sensitivity = 0
if float(confusion[1,1]+confusion[1,0])!=0:
sensitivity = float(confusion[1,1])/float(confusion[1,1]+confusion[1,0])
print("Sensitivity: " +str(sensitivity))
precision = 0
if float(confusion[1,1]+confusion[0,1])!=0:
precision = float(confusion[1,1])/float(confusion[1,1]+confusion[0,1])
print("Precision: " +str(precision))
#Jaccard similarity index
jaccard_index = jaccard_similarity_score(y_true, y_pred, normalize=True)
print("\nJaccard similarity score: " +str(jaccard_index))
#F1 score
F1_score = f1_score(y_true, y_pred, labels=None, average='binary', sample_weight=None)
print("\nF1 score (F-measure): " +str(F1_score))
#Save the results
file_perf = open(path_experiment+'performances.txt', 'w')
file_perf.write("Area under the ROC curve: "+str(AUC_ROC)
+ "\nArea under Precision-Recall curve: " +str(AUC_prec_rec)
+ "\nJaccard similarity score: " +str(jaccard_index)
+ "\nF1 score (F-measure): " +str(F1_score)
+"\n\nConfusion matrix:"
+str(confusion)
+"\nACCURACY: " +str(accuracy)
+"\nSENSITIVITY: " +str(sensitivity)
+"\nSPECIFICITY: " +str(specificity)
+"\nPRECISION: " +str(precision)
)
file_perf.close()
================================================
FILE: src/retinaNN_training.py
================================================
###################################################
#
# Script to:
# - Load the images and extract the patches
# - Define the neural network
# - define the training
#
##################################################
import numpy as np
from six.moves import configparser
import torch.backends.cudnn as cudnn
import sys
sys.path.insert(0, '../')
from lib.help_functions import *
#function to obtain data for training/testing (validation)
from lib.extract_patches import get_data_training
import os
from losses import *
import torch.optim as optim
from torch.utils.data import Dataset, DataLoader
from tqdm import tqdm
import random
def count_parameters(model):
return sum(p.numel() for p in model.parameters() if p.requires_grad)
#========= Load settings from Config file
config = configparser.RawConfigParser()
config.read('../configuration.txt')
#patch to the datasets
path_data = config.get('data paths', 'path_local')
#Experiment name
name_experiment = config.get('experiment name', 'name')
#training settings
N_epochs = int(config.get('training settings', 'N_epochs'))
batch_size = int(config.get('training settings', 'batch_size'))
#========== Define parameters here =============================
# log file
if not os.path.exists('./logs'):
os.mkdir('logs')
device = 'cuda' if torch.cuda.is_available() else 'cpu'
start_epoch = 0 # start from epoch 0 or last checkpoint epoch
total_epoch = 200
val_portion = 0.1
lr_epoch = np.array([150,total_epoch])
lr_value= np.array([0.001,0.0001])
layers = 4
filters = 10
from LadderNetv65 import LadderNetv6
net = LadderNetv6(num_classes=2,layers=layers,filters=filters,inplanes=1)
print("Toral number of parameters: "+str(count_parameters(net)))
check_path = 'LadderNetv65_layer_%d_filter_%d.pt7'% (layers,filters) #'UNet16.pt7'#'UNet_Resnet101.pt7'
resume = False
criterion = LossMulti(jaccard_weight=0)
#criterion = CrossEntropy2d()
#optimizer = optim.SGD(net.parameters(),
# lr=lr_schedule[0], momentum=0.9, weight_decay=5e-4, nesterov=True)
optimizer = optim.Adam(net.parameters(),lr=lr_value[0])
#============ Load the data and divided in patches
patches_imgs_train, patches_masks_train = get_data_training(
DRIVE_train_imgs_original = path_data + config.get('data paths', 'train_imgs_original'),
DRIVE_train_groudTruth = path_data + config.get('data paths', 'train_groundTruth'), #masks
patch_height = int(config.get('data attributes', 'patch_height')),
patch_width = int(config.get('data attributes', 'patch_width')),
N_subimgs = int(config.get('training settings', 'N_subimgs')),
inside_FOV = config.getboolean('training settings', 'inside_FOV') #select the patches only inside the FOV (default == True)
)
class TrainDataset(Dataset):
"""Endovis 2018 dataset."""
def __init__(self, patches_imgs,patches_masks_train):
self.imgs = patches_imgs
self.masks = patches_masks_train
def __len__(self):
return self.imgs.shape[0]
def __getitem__(self, idx):
tmp = self.masks[idx]
tmp = np.squeeze(tmp,0)
return torch.from_numpy(self.imgs[idx,...]).float(), torch.from_numpy(tmp).long()
val_ind = random.sample(range(patches_masks_train.shape[0]),int(np.floor(val_portion*patches_masks_train.shape[0])))
train_ind = set(range(patches_masks_train.shape[0])) - set(val_ind)
train_ind = list(train_ind)
train_set = TrainDataset(patches_imgs_train[train_ind,...],patches_masks_train[train_ind,...])
train_loader = DataLoader(train_set, batch_size=batch_size,
shuffle=True, num_workers=4)
val_set = TrainDataset(patches_imgs_train[val_ind,...],patches_masks_train[val_ind,...])
val_loader = DataLoader(val_set, batch_size=batch_size,
shuffle=True, num_workers=4)
#========= Save a sample of what you're feeding to the neural network ==========
N_sample = min(patches_imgs_train.shape[0],40)
visualize(group_images(patches_imgs_train[0:N_sample,:,:,:],5),'../'+name_experiment+'/'+"sample_input_imgs")#.show()
visualize(group_images(patches_masks_train[0:N_sample,:,:,:],5),'../'+name_experiment+'/'+"sample_input_masks")#.show()
best_loss = np.Inf
# create a list of learning rate with epochs
lr_schedule = np.zeros(total_epoch)
for l in range(len(lr_epoch)):
if l ==0:
lr_schedule[0:lr_epoch[l]] = lr_value[l]
else:
lr_schedule[lr_epoch[l-1]:lr_epoch[l]] = lr_value[l]
if device == 'cuda':
net.cuda()
net = torch.nn.DataParallel(net, device_ids=range(torch.cuda.device_count()))
cudnn.benchmark = True
if resume:
# Load checkpoint.
print('==> Resuming from checkpoint..')
assert os.path.isdir('checkpoint'), 'Error: no checkpoint directory found!'
checkpoint = torch.load('./checkpoint/'+check_path)
net.load_state_dict(checkpoint['net'])
start_epoch = checkpoint['epoch']
def train(epoch):
print('\nEpoch: %d' % epoch)
net.train()
train_loss = 0
IoU = []
# get learning rate from learing schedule
lr = lr_schedule[epoch]
for param_group in optimizer.param_groups:
param_group['lr'] = lr
print("Learning rate = %4f\n" % lr)
IU = []
# train network
for batch_idx, (inputs, targets) in enumerate(tqdm(train_loader)):
inputs, targets = inputs.to(device), targets.to(device)
optimizer.zero_grad()
outputs = net(inputs)
loss = criterion(outputs, targets)
loss.backward()
optimizer.step()
train_loss += loss.item()
print("Epoch %d: Train loss %4f\n" % (epoch, train_loss / np.float32(len(train_loader))))
def test(epoch, display=False):
global best_loss
net.eval()
test_loss = 0
with torch.no_grad():
for batch_idx, (inputs, targets) in enumerate(val_loader):
inputs, targets = inputs.to(device), targets.to(device)
outputs = net(inputs)
loss = criterion(outputs, targets)
test_loss += loss.item()
print(
'Valid loss: {:.4f}'.format(test_loss))
# Save checkpoint.
if test_loss < best_loss:
print('Saving..')
state = {
'net': net.state_dict(),
'best_loss': best_loss,
'epoch': epoch,
}
if not os.path.isdir('checkpoint'):
os.mkdir('checkpoint')
torch.save(state, './checkpoint/' + check_path)
best_loss = test_loss
for epoch in range(start_epoch,total_epoch):
train(epoch)
test(epoch,False)
================================================
FILE: src/subpixel_upscaling.py
================================================
# -*- coding: utf-8 -*-
"""
This file contains an implementation of Sub-pixel convolutional upscaling layer based on
the paper "Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel
Convolutional Neural Network" (https://arxiv.org/abs/1609.05158).
"""
from __future__ import absolute_import
import tensorflow as tf
from keras.engine import Layer
from keras.utils.generic_utils import get_custom_objects
from keras.utils.conv_utils import normalize_data_format
class SubPixelUpscaling(Layer):
""" This layer requires a Convolution2D prior to it, having output filters computed according to
the formula :
filters = k * (scale_factor * scale_factor)
where k = a user defined number of filters (generally larger than 32)
scale_factor = the upscaling factor (generally 2)
This layer performs the depth to space operation on the convolution filters, and returns a
tensor with the size as defined below.
# Example :
```python
# A standard subpixel upscaling block
x = Convolution2D(256, 3, 3, padding='same', activation='relu')(...)
u = SubPixelUpscaling(scale_factor=2)(x)
[Optional]
x = Convolution2D(256, 3, 3, padding='same', activation='relu')(u)
```
In practice, it is useful to have a second convolution layer after the
SubPixelUpscaling layer to speed up the learning process.
However, if you are stacking multiple SubPixelUpscaling blocks, it may increase
the number of parameters greatly, so the Convolution layer after SubPixelUpscaling
layer can be removed.
# Arguments
scale_factor: Upscaling factor.
data_format: Can be None, 'channels_first' or 'channels_last'.
# Input shape
4D tensor with shape:
`(samples, k * (scale_factor * scale_factor) channels, rows, cols)` if data_format='channels_first'
or 4D tensor with shape:
`(samples, rows, cols, k * (scale_factor * scale_factor) channels)` if data_format='channels_last'.
# Output shape
4D tensor with shape:
`(samples, k channels, rows * scale_factor, cols * scale_factor))` if data_format='channels_first'
or 4D tensor with shape:
`(samples, rows * scale_factor, cols * scale_factor, k channels)` if data_format='channels_last'.
"""
def __init__(self, scale_factor=2, data_format=None, **kwargs):
super(SubPixelUpscaling, self).__init__(**kwargs)
self.scale_factor = scale_factor
self.data_format = normalize_data_format(data_format)
def build(self, input_shape):
pass
def call(self, x, mask=None):
y = tf.depth_to_space(x, self.scale_factor, self.data_format)
return y
def compute_output_shape(self, input_shape):
if self.data_format == 'channels_first':
b, k, r, c = input_shape
return (b, k // (self.scale_factor ** 2), r * self.scale_factor, c * self.scale_factor)
else:
b, r, c, k = input_shape
return (b, r * self.scale_factor, c * self.scale_factor, k // (self.scale_factor ** 2))
def get_config(self):
config = {'scale_factor': self.scale_factor,
'data_format': self.data_format}
base_config = super(SubPixelUpscaling, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
get_custom_objects().update({'SubPixelUpscaling': SubPixelUpscaling})
================================================
FILE: test/drive_performances.txt
================================================
Area under the ROC curve: 0.9793486169167707
Area under Precision-Recall curve: 0.9107415769226219
Jaccard similarity score: 0.9561483628876393
F1 score (F-measure): 0.8201609835049293
Confusion matrix:[[3885354 75140]
[ 123865 453784]]
ACCURACY: 0.9561483628876393
SENSITIVITY: 0.7855704761888275
SPECIFICITY: 0.981027619281837
PRECISION: 0.8579380024351324
================================================
FILE: test/performances.txt
================================================
Area under the ROC curve: 0.9793257409827817
Area under Precision-Recall curve: 0.9107420271043044
Jaccard similarity score: 0.9561715001047786
F1 score (F-measure): 0.8202472616852836
Confusion matrix:[[3885433 75061]
[ 123839 453810]]
ACCURACY: 0.9561715001047786
SENSITIVITY: 0.7856154862208712
SPECIFICITY: 0.9810475662884478
PRECISION: 0.8580731407091711
================================================
FILE: test/test_architecture.json
================================================
{"class_name": "Model", "config": {"name": "model_1", "layers": [{"name": "input_1", "class_name": "InputLayer", "config": {"batch_input_shape": [null, 48, 48, 1], "dtype": "float32", "sparse": false, "name": "input_1"}, "inbound_nodes": []}, {"name": "conv2d_1", "class_name": "Conv2D", "config": {"name": "conv2d_1", "trainable": true, "filters": 10, "kernel_size": [3, 3], "strides": [1, 1], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["input_1", 0, 0, {}]]]}, {"name": "conv2d_2", "class_name": "Conv2D", "config": {"name": "conv2d_2", "trainable": true, "filters": 10, "kernel_size": [3, 3], "strides": [1, 1], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_1", 0, 0, {}]], [["dropout_1", 0, 0, {}]], [["add_1", 0, 0, {}]]]}, {"name": "dropout_1", "class_name": "Dropout", "config": {"name": "dropout_1", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_2", 0, 0, {}]]]}, {"name": "dropout_2", "class_name": "Dropout", "config": {"name": "dropout_2", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_2", 1, 0, {}]]]}, {"name": "add_1", "class_name": "Add", "config": {"name": "add_1", "trainable": true}, "inbound_nodes": [[["conv2d_1", 0, 0, {}], ["dropout_2", 0, 0, {}]]]}, {"name": "conv2d_3", "class_name": "Conv2D", "config": {"name": "conv2d_3", "trainable": true, "filters": 20, "kernel_size": [3, 3], "strides": [2, 2], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_2", 2, 0, {}]]]}, {"name": "conv2d_4", "class_name": "Conv2D", "config": {"name": "conv2d_4", "trainable": true, "filters": 20, "kernel_size": [3, 3], "strides": [1, 1], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_3", 0, 0, {}]], [["dropout_3", 0, 0, {}]], [["add_2", 0, 0, {}]]]}, {"name": "dropout_3", "class_name": "Dropout", "config": {"name": "dropout_3", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_4", 0, 0, {}]]]}, {"name": "dropout_4", "class_name": "Dropout", "config": {"name": "dropout_4", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_4", 1, 0, {}]]]}, {"name": "add_2", "class_name": "Add", "config": {"name": "add_2", "trainable": true}, "inbound_nodes": [[["conv2d_3", 0, 0, {}], ["dropout_4", 0, 0, {}]]]}, {"name": "conv2d_5", "class_name": "Conv2D", "config": {"name": "conv2d_5", "trainable": true, "filters": 40, "kernel_size": [3, 3], "strides": [2, 2], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_4", 2, 0, {}]]]}, {"name": "conv2d_6", "class_name": "Conv2D", "config": {"name": "conv2d_6", "trainable": true, "filters": 40, "kernel_size": [3, 3], "strides": [1, 1], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_5", 0, 0, {}]], [["dropout_5", 0, 0, {}]], [["add_3", 0, 0, {}]]]}, {"name": "dropout_5", "class_name": "Dropout", "config": {"name": "dropout_5", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_6", 0, 0, {}]]]}, {"name": "dropout_6", "class_name": "Dropout", "config": {"name": "dropout_6", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_6", 1, 0, {}]]]}, {"name": "add_3", "class_name": "Add", "config": {"name": "add_3", "trainable": true}, "inbound_nodes": [[["conv2d_5", 0, 0, {}], ["dropout_6", 0, 0, {}]]]}, {"name": "conv2d_7", "class_name": "Conv2D", "config": {"name": "conv2d_7", "trainable": true, "filters": 80, "kernel_size": [3, 3], "strides": [2, 2], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_6", 2, 0, {}]]]}, {"name": "conv2d_8", "class_name": "Conv2D", "config": {"name": "conv2d_8", "trainable": true, "filters": 80, "kernel_size": [3, 3], "strides": [1, 1], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_7", 0, 0, {}]], [["dropout_7", 0, 0, {}]], [["add_4", 0, 0, {}]]]}, {"name": "dropout_7", "class_name": "Dropout", "config": {"name": "dropout_7", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_8", 0, 0, {}]]]}, {"name": "dropout_8", "class_name": "Dropout", "config": {"name": "dropout_8", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_8", 1, 0, {}]]]}, {"name": "add_4", "class_name": "Add", "config": {"name": "add_4", "trainable": true}, "inbound_nodes": [[["conv2d_7", 0, 0, {}], ["dropout_8", 0, 0, {}]]]}, {"name": "conv2d_9", "class_name": "Conv2D", "config": {"name": "conv2d_9", "trainable": true, "filters": 160, "kernel_size": [3, 3], "strides": [2, 2], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_8", 2, 0, {}]]]}, {"name": "conv2d_10", "class_name": "Conv2D", "config": {"name": "conv2d_10", "trainable": true, "filters": 160, "kernel_size": [3, 3], "strides": [1, 1], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_9", 0, 0, {}]], [["dropout_9", 0, 0, {}]], [["add_5", 0, 0, {}]]]}, {"name": "dropout_9", "class_name": "Dropout", "config": {"name": "dropout_9", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_10", 0, 0, {}]]]}, {"name": "dropout_10", "class_name": "Dropout", "config": {"name": "dropout_10", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_10", 1, 0, {}]]]}, {"name": "add_5", "class_name": "Add", "config": {"name": "add_5", "trainable": true}, "inbound_nodes": [[["conv2d_9", 0, 0, {}], ["dropout_10", 0, 0, {}]]]}, {"name": "conv2d_transpose_1", "class_name": "Conv2DTranspose", "config": {"name": "conv2d_transpose_1", "trainable": true, "filters": 80, "kernel_size": [3, 3], "strides": [2, 2], "padding": "same", "data_format": "channels_last", "activation": "linear", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_10", 2, 0, {}]]]}, {"name": "add_6", "class_name": "Add", "config": {"name": "add_6", "trainable": true}, "inbound_nodes": [[["conv2d_transpose_1", 0, 0, {}], ["conv2d_8", 2, 0, {}]]]}, {"name": "conv2d_11", "class_name": "Conv2D", "config": {"name": "conv2d_11", "trainable": true, "filters": 80, "kernel_size": [3, 3], "strides": [1, 1], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["add_6", 0, 0, {}]], [["dropout_11", 0, 0, {}]], [["add_7", 0, 0, {}]]]}, {"name": "dropout_11", "class_name": "Dropout", "config": {"name": "dropout_11", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_11", 0, 0, {}]]]}, {"name": "dropout_12", "class_name": "Dropout", "config": {"name": "dropout_12", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_11", 1, 0, {}]]]}, {"name": "add_7", "class_name": "Add", "config": {"name": "add_7", "trainable": true}, "inbound_nodes": [[["add_6", 0, 0, {}], ["dropout_12", 0, 0, {}]]]}, {"name": "conv2d_transpose_2", "class_name": "Conv2DTranspose", "config": {"name": "conv2d_transpose_2", "trainable": true, "filters": 40, "kernel_size": [3, 3], "strides": [2, 2], "padding": "same", "data_format": "channels_last", "activation": "linear", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_11", 2, 0, {}]]]}, {"name": "add_8", "class_name": "Add", "config": {"name": "add_8", "trainable": true}, "inbound_nodes": [[["conv2d_transpose_2", 0, 0, {}], ["conv2d_6", 2, 0, {}]]]}, {"name": "conv2d_12", "class_name": "Conv2D", "config": {"name": "conv2d_12", "trainable": true, "filters": 40, "kernel_size": [3, 3], "strides": [1, 1], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["add_8", 0, 0, {}]], [["dropout_13", 0, 0, {}]], [["add_9", 0, 0, {}]]]}, {"name": "dropout_13", "class_name": "Dropout", "config": {"name": "dropout_13", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_12", 0, 0, {}]]]}, {"name": "dropout_14", "class_name": "Dropout", "config": {"name": "dropout_14", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_12", 1, 0, {}]]]}, {"name": "add_9", "class_name": "Add", "config": {"name": "add_9", "trainable": true}, "inbound_nodes": [[["add_8", 0, 0, {}], ["dropout_14", 0, 0, {}]]]}, {"name": "conv2d_transpose_3", "class_name": "Conv2DTranspose", "config": {"name": "conv2d_transpose_3", "trainable": true, "filters": 20, "kernel_size": [3, 3], "strides": [2, 2], "padding": "same", "data_format": "channels_last", "activation": "linear", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_12", 2, 0, {}]]]}, {"name": "add_10", "class_name": "Add", "config": {"name": "add_10", "trainable": true}, "inbound_nodes": [[["conv2d_transpose_3", 0, 0, {}], ["conv2d_4", 2, 0, {}]]]}, {"name": "conv2d_13", "class_name": "Conv2D", "config": {"name": "conv2d_13", "trainable": true, "filters": 20, "kernel_size": [3, 3], "strides": [1, 1], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["add_10", 0, 0, {}]], [["dropout_15", 0, 0, {}]], [["add_11", 0, 0, {}]]]}, {"name": "dropout_15", "class_name": "Dropout", "config": {"name": "dropout_15", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_13", 0, 0, {}]]]}, {"name": "dropout_16", "class_name": "Dropout", "config": {"name": "dropout_16", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_13", 1, 0, {}]]]}, {"name": "add_11", "class_name": "Add", "config": {"name": "add_11", "trainable": true}, "inbound_nodes": [[["add_10", 0, 0, {}], ["dropout_16", 0, 0, {}]]]}, {"name": "conv2d_transpose_4", "class_name": "Conv2DTranspose", "config": {"name": "conv2d_transpose_4", "trainable": true, "filters": 10, "kernel_size": [3, 3], "strides": [2, 2], "padding": "same", "data_format": "channels_last", "activation": "linear", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_13", 2, 0, {}]]]}, {"name": "add_12", "class_name": "Add", "config": {"name": "add_12", "trainable": true}, "inbound_nodes": [[["conv2d_transpose_4", 0, 0, {}], ["conv2d_2", 2, 0, {}]]]}, {"name": "conv2d_14", "class_name": "Conv2D", "config": {"name": "conv2d_14", "trainable": true, "filters": 10, "kernel_size": [3, 3], "strides": [1, 1], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["add_12", 0, 0, {}]], [["dropout_17", 0, 0, {}]], [["add_13", 0, 0, {}]]]}, {"name": "dropout_17", "class_name": "Dropout", "config": {"name": "dropout_17", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_14", 0, 0, {}]]]}, {"name": "dropout_18", "class_name": "Dropout", "config": {"name": "dropout_18", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_14", 1, 0, {}]]]}, {"name": "add_13", "class_name": "Add", "config": {"name": "add_13", "trainable": true}, "inbound_nodes": [[["add_12", 0, 0, {}], ["dropout_18", 0, 0, {}]]]}, {"name": "conv2d_15", "class_name": "Conv2D", "config": {"name": "conv2d_15", "trainable": true, "filters": 10, "kernel_size": [3, 3], "strides": [1, 1], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_14", 2, 0, {}]]]}, {"name": "add_14", "class_name": "Add", "config": {"name": "add_14", "trainable": true}, "inbound_nodes": [[["conv2d_15", 0, 0, {}], ["conv2d_14", 2, 0, {}]]]}, {"name": "conv2d_16", "class_name": "Conv2D", "config": {"name": "conv2d_16", "trainable": true, "filters": 10, "kernel_size": [3, 3], "strides": [1, 1], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["add_14", 0, 0, {}]], [["dropout_19", 0, 0, {}]], [["add_15", 0, 0, {}]]]}, {"name": "dropout_19", "class_name": "Dropout", "config": {"name": "dropout_19", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_16", 0, 0, {}]]]}, {"name": "dropout_20", "class_name": "Dropout", "config": {"name": "dropout_20", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_16", 1, 0, {}]]]}, {"name": "add_15", "class_name": "Add", "config": {"name": "add_15", "trainable": true}, "inbound_nodes": [[["add_14", 0, 0, {}], ["dropout_20", 0, 0, {}]]]}, {"name": "conv2d_17", "class_name": "Conv2D", "config": {"name": "conv2d_17", "trainable": true, "filters": 20, "kernel_size": [3, 3], "strides": [2, 2], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_16", 2, 0, {}]]]}, {"name": "add_16", "class_name": "Add", "config": {"name": "add_16", "trainable": true}, "inbound_nodes": [[["conv2d_17", 0, 0, {}], ["conv2d_13", 2, 0, {}]]]}, {"name": "conv2d_18", "class_name": "Conv2D", "config": {"name": "conv2d_18", "trainable": true, "filters": 20, "kernel_size": [3, 3], "strides": [1, 1], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["add_16", 0, 0, {}]], [["dropout_21", 0, 0, {}]], [["add_17", 0, 0, {}]]]}, {"name": "dropout_21", "class_name": "Dropout", "config": {"name": "dropout_21", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_18", 0, 0, {}]]]}, {"name": "dropout_22", "class_name": "Dropout", "config": {"name": "dropout_22", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_18", 1, 0, {}]]]}, {"name": "add_17", "class_name": "Add", "config": {"name": "add_17", "trainable": true}, "inbound_nodes": [[["add_16", 0, 0, {}], ["dropout_22", 0, 0, {}]]]}, {"name": "conv2d_19", "class_name": "Conv2D", "config": {"name": "conv2d_19", "trainable": true, "filters": 40, "kernel_size": [3, 3], "strides": [2, 2], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["conv2d_18", 2, 0, {}]]]}, {"name": "add_18", "class_name": "Add", "config": {"name": "add_18", "trainable": true}, "inbound_nodes": [[["conv2d_19", 0, 0, {}], ["conv2d_12", 2, 0, {}]]]}, {"name": "conv2d_20", "class_name": "Conv2D", "config": {"name": "conv2d_20", "trainable": true, "filters": 40, "kernel_size": [3, 3], "strides": [1, 1], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}}, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null}, "inbound_nodes": [[["add_18", 0, 0, {}]], [["dropout_23", 0, 0, {}]], [["add_19", 0, 0, {}]]]}, {"name": "dropout_23", "class_name": "Dropout", "config": {"name": "dropout_23", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_20", 0, 0, {}]]]}, {"name": "dropout_24", "class_name": "Dropout", "config": {"name": "dropout_24", "trainable": true, "rate": 0.25, "noise_shape": null, "seed": null}, "inbound_nodes": [[["conv2d_20", 1, 0, {}]]]}, {"name": "add_19", "class_name": "Add", "config": {"name": "add_19", "trainable": true}, "inbound_nodes": [[["add_18", 0, 0, {}], ["dropout_24", 0, 0, {}]]]}, {"name": "conv2d_21", "class_name": "Conv2D", "config": {"name": "conv2d_21", "trainable": true, "filters": 80, "kernel_size": [3, 3], "strides": [2, 2], "padding": "same", "data_format": "channels_last", "dilation_rate": [1, 1], "activation": "relu", "use_bias": true, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"scale": 1.0, "mode": "fan_avg", "distribution": "uniform", "seed": null}}, "bias_initializer": {"class_name": "Zeros", "config": {}},
gitextract_5b2xktow/
├── .gitignore
├── .idea/
│ ├── libraries/
│ │ └── R_User_Library.xml
│ ├── misc.xml
│ ├── modules.xml
│ ├── retinaunet-master.iml
│ └── workspace.xml
├── DRIVE/
│ ├── test/
│ │ └── images/
│ │ ├── 01_test.tif
│ │ ├── 02_test.tif
│ │ ├── 03_test.tif
│ │ ├── 04_test.tif
│ │ ├── 05_test.tif
│ │ ├── 06_test.tif
│ │ ├── 07_test.tif
│ │ ├── 08_test.tif
│ │ ├── 09_test.tif
│ │ ├── 10_test.tif
│ │ ├── 11_test.tif
│ │ ├── 12_test.tif
│ │ ├── 13_test.tif
│ │ ├── 14_test.tif
│ │ ├── 15_test.tif
│ │ ├── 16_test.tif
│ │ ├── 17_test.tif
│ │ ├── 18_test.tif
│ │ ├── 19_test.tif
│ │ └── 20_test.tif
│ └── training/
│ └── images/
│ ├── 21_training.tif
│ ├── 22_training.tif
│ ├── 23_training.tif
│ ├── 24_training.tif
│ ├── 25_training.tif
│ ├── 26_training.tif
│ ├── 27_training.tif
│ ├── 28_training.tif
│ ├── 29_training.tif
│ ├── 30_training.tif
│ ├── 31_training.tif
│ ├── 32_training.tif
│ ├── 33_training.tif
│ ├── 34_training.tif
│ ├── 35_training.tif
│ ├── 36_training.tif
│ ├── 37_training.tif
│ ├── 38_training.tif
│ ├── 39_training.tif
│ └── 40_training.tif
├── README.md
├── __init__.py
├── _config.yml
├── configuration.txt
├── figures/
│ └── readme
├── googlefffa84c6c8a268c9.html
├── lib/
│ ├── __init__.py
│ ├── densenet.py
│ ├── extract_patches.py
│ ├── extract_patches.py.bak
│ ├── help_functions.py
│ ├── help_functions.py.bak
│ ├── pre_processing.py
│ └── pre_processing.py.bak
├── prepare_datasets_DRIVE.py
├── src/
│ ├── LadderNetv65.py
│ ├── __init__.py
│ ├── checkpoint/
│ │ └── LadderNetv65_layer_4_filter_10.pt7
│ ├── losses.py
│ ├── retinaNN_predict.py
│ ├── retinaNN_training.py
│ └── subpixel_upscaling.py
└── test/
├── drive_performances.txt
├── performances.txt
├── test_architecture.json
├── test_best_weights.h5
├── test_configuration.txt
├── test_last_weights.h5
└── test_training.nohup
SYMBOL INDEX (84 symbols across 10 files)
FILE: lib/densenet.py
function preprocess_input (line 90) | def preprocess_input(x, data_format=None):
function DenseNet (line 105) | def DenseNet(input_shape=None,
function DenseNetFCN (line 307) | def DenseNetFCN(input_shape, nb_dense_block=5, growth_rate=16, nb_layers...
function DenseNetImageNet121 (line 436) | def DenseNetImageNet121(input_shape=None,
function DenseNetImageNet169 (line 454) | def DenseNetImageNet169(input_shape=None,
function DenseNetImageNet201 (line 472) | def DenseNetImageNet201(input_shape=None,
function DenseNetImageNet264 (line 490) | def DenseNetImageNet264(input_shape=None,
function DenseNetImageNet161 (line 508) | def DenseNetImageNet161(input_shape=None,
function name_or_none (line 526) | def name_or_none(prefix, name):
function __conv_block (line 530) | def __conv_block(ip, nb_filter, bottleneck=False, dropout_rate=None, wei...
function __dense_block (line 583) | def __dense_block(x, nb_layers, nb_filter, growth_rate, bottleneck=False...
function __transition_block (line 633) | def __transition_block(ip, nb_filter, compression=1.0, weight_decay=1e-4...
function __transition_up_block (line 680) | def __transition_up_block(ip, nb_filters, type='deconv', weight_decay=1E...
function __create_dense_net (line 724) | def __create_dense_net(nb_classes, img_input, include_top, depth=40, nb_...
function __create_fcn_dense_net (line 864) | def __create_fcn_dense_net(nb_classes, img_input, include_top, nb_dense_...
FILE: lib/extract_patches.py
function get_data_training (line 16) | def get_data_training(DRIVE_train_imgs_original,
function get_data_testing (line 54) | def get_data_testing(DRIVE_test_imgs_original, DRIVE_test_groudTruth, Im...
function get_data_testing_overlap (line 94) | def get_data_testing_overlap(DRIVE_test_imgs_original, DRIVE_test_groudT...
function data_consistency_check (line 127) | def data_consistency_check(imgs,masks):
function extract_random (line 138) | def extract_random(full_imgs,full_masks, patch_h,patch_w, N_patches, ins...
function is_patch_inside_FOV (line 175) | def is_patch_inside_FOV(x,y,img_w,img_h,patch_h):
function extract_ordered (line 187) | def extract_ordered(full_imgs, patch_h, patch_w):
function paint_border_overlap (line 213) | def paint_border_overlap(full_imgs, patch_h, patch_w, stride_h, stride_w):
function extract_ordered_overlap (line 240) | def extract_ordered_overlap(full_imgs, patch_h, patch_w,stride_h,stride_w):
function recompone_overlap (line 263) | def recompone_overlap(preds, img_h, img_w, stride_h, stride_w):
function recompone (line 297) | def recompone(data,N_h,N_w):
function paint_border (line 324) | def paint_border(data,patch_h,patch_w):
function pred_only_FOV (line 345) | def pred_only_FOV(data_imgs,data_masks,original_imgs_border_masks):
function kill_border (line 366) | def kill_border(data, original_imgs_border_masks):
function inside_FOV_DRIVE (line 378) | def inside_FOV_DRIVE(i, x, y, DRIVE_masks):
FILE: lib/help_functions.py
function load_hdf5 (line 6) | def load_hdf5(infile):
function write_hdf5 (line 10) | def write_hdf5(arr,outfile):
function rgb2gray (line 15) | def rgb2gray(rgb):
function group_images (line 23) | def group_images(data,per_row):
function visualize (line 40) | def visualize(data,filename):
function masks_Unet (line 54) | def masks_Unet(masks):
function pred_to_imgs (line 76) | def pred_to_imgs(pred, patch_height, patch_width, mode="original"):
FILE: lib/pre_processing.py
function my_PreProc (line 16) | def my_PreProc(data):
function histo_equalized (line 34) | def histo_equalized(imgs):
function clahe_equalized (line 45) | def clahe_equalized(imgs):
function dataset_normalized (line 57) | def dataset_normalized(imgs):
function adjust_gamma (line 69) | def adjust_gamma(imgs, gamma=1.0):
FILE: prepare_datasets_DRIVE.py
function write_hdf5 (line 14) | def write_hdf5(arr,outfile):
function get_datasets (line 39) | def get_datasets(imgs_dir,groundTruth_dir,borderMasks_dir,train_test="nu...
FILE: src/LadderNetv65.py
function conv3x3 (line 7) | def conv3x3(in_planes, out_planes, stride=1):
class BasicBlock (line 13) | class BasicBlock(nn.Module):
method __init__ (line 16) | def __init__(self, inplanes, planes, stride=1, downsample=None):
method forward (line 33) | def forward(self, x):
class Bottleneck (line 52) | class Bottleneck(nn.Module):
method __init__ (line 55) | def __init__(self, inplanes, planes, stride=1, downsample=None):
method forward (line 68) | def forward(self, x):
class Initial_LadderBlock (line 90) | class Initial_LadderBlock(nn.Module):
method __init__ (line 92) | def __init__(self,planes,layers,kernel=3,block=BasicBlock,inplanes = 3):
method forward (line 124) | def forward(self, x):
class LadderBlock (line 152) | class LadderBlock(nn.Module):
method __init__ (line 154) | def __init__(self,planes,layers,kernel=3,block=BasicBlock,inplanes = 3):
method forward (line 185) | def forward(self, x):
class Final_LadderBlock (line 214) | class Final_LadderBlock(nn.Module):
method __init__ (line 216) | def __init__(self,planes,layers,kernel=3,block=BasicBlock,inplanes = 3):
method forward (line 220) | def forward(self, x):
class LadderNetv6 (line 224) | class LadderNetv6(nn.Module):
method __init__ (line 225) | def __init__(self,layers=3,filters=16,num_classes=2,inplanes=3):
method forward (line 232) | def forward(self,x):
FILE: src/losses.py
function cuda (line 5) | def cuda(x):
class LossMulti (line 9) | class LossMulti:
method __init__ (line 10) | def __init__(self, jaccard_weight=0, class_weights=None, num_classes=1):
method __call__ (line 20) | def __call__(self, outputs, targets):
FILE: src/retinaNN_predict.py
class TrainDataset (line 147) | class TrainDataset(Dataset):
method __init__ (line 150) | def __init__(self, patches_imgs):
method __len__ (line 153) | def __len__(self):
method __getitem__ (line 156) | def __getitem__(self, idx):
FILE: src/retinaNN_training.py
function count_parameters (line 34) | def count_parameters(model):
class TrainDataset (line 94) | class TrainDataset(Dataset):
method __init__ (line 97) | def __init__(self, patches_imgs,patches_masks_train):
method __len__ (line 101) | def __len__(self):
method __getitem__ (line 104) | def __getitem__(self, idx):
function train (line 149) | def train(epoch):
function test (line 177) | def test(epoch, display=False):
FILE: src/subpixel_upscaling.py
class SubPixelUpscaling (line 15) | class SubPixelUpscaling(Layer):
method __init__ (line 51) | def __init__(self, scale_factor=2, data_format=None, **kwargs):
method build (line 57) | def build(self, input_shape):
method call (line 60) | def call(self, x, mask=None):
method compute_output_shape (line 64) | def compute_output_shape(self, input_shape):
method get_config (line 72) | def get_config(self):
Condensed preview — 75 files, each showing path, character count, and a content snippet. Download the .json file or copy for the full structured content (223K chars).
[
{
"path": ".gitignore",
"chars": 50,
"preview": "*~\n*.pyc\n.DS_Store\nDRIVE_datasets_training_testing"
},
{
"path": ".idea/libraries/R_User_Library.xml",
"chars": 123,
"preview": "<component name=\"libraryTable\">\n <library name=\"R User Library\">\n <CLASSES />\n <SOURCES />\n </library>\n</compone"
},
{
"path": ".idea/misc.xml",
"chars": 198,
"preview": "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<project version=\"4\">\n <component name=\"ProjectRootManager\" version=\"2\" project-"
},
{
"path": ".idea/modules.xml",
"chars": 286,
"preview": "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<project version=\"4\">\n <component name=\"ProjectModuleManager\">\n <modules>\n "
},
{
"path": ".idea/retinaunet-master.iml",
"chars": 746,
"preview": "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<module type=\"PYTHON_MODULE\" version=\"4\">\n <component name=\"NewModuleRootManager"
},
{
"path": ".idea/workspace.xml",
"chars": 29358,
"preview": "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<project version=\"4\">\n <component name=\"BookmarkManager\">\n <bookmark url=\"fil"
},
{
"path": "README.md",
"chars": 3053,
"preview": "* Implementation for paper [LadderNet: Multi-path networks based on U-Net for medical image segmentation\n](https://arxiv"
},
{
"path": "__init__.py",
"chars": 0,
"preview": ""
},
{
"path": "_config.yml",
"chars": 26,
"preview": "theme: jekyll-theme-cayman"
},
{
"path": "configuration.txt",
"chars": 1344,
"preview": "[data paths]\npath_local = ../DRIVE_datasets_training_testing/\ntrain_imgs_original = DRIVE_dataset_imgs_train.hdf5\ntrain"
},
{
"path": "figures/readme",
"chars": 5,
"preview": "test\n"
},
{
"path": "googlefffa84c6c8a268c9.html",
"chars": 53,
"preview": "google-site-verification: googlefffa84c6c8a268c9.html"
},
{
"path": "lib/__init__.py",
"chars": 0,
"preview": ""
},
{
"path": "lib/densenet.py",
"chars": 49534,
"preview": "# -*- coding: utf-8 -*-\n'''\ncode from Keras.contrib. This is just a local copy in case that repo changes.\n\nDenseNet and "
},
{
"path": "lib/extract_patches.py",
"chars": 18473,
"preview": "import numpy as np\nimport random\nimport configparser\n\nfrom .help_functions import load_hdf5\nfrom .help_functions import "
},
{
"path": "lib/extract_patches.py.bak",
"chars": 18424,
"preview": "import numpy as np\nimport random\nimport ConfigParser\n\nfrom help_functions import load_hdf5\nfrom help_functions import vi"
},
{
"path": "lib/help_functions.py",
"chars": 3425,
"preview": "import h5py\nimport numpy as np\nfrom PIL import Image\nfrom matplotlib import pyplot as plt\n\ndef load_hdf5(infile):\n with"
},
{
"path": "lib/help_functions.py.bak",
"chars": 3344,
"preview": "import h5py\nimport numpy as np\nfrom PIL import Image\nfrom matplotlib import pyplot as plt\n\ndef load_hdf5(infile):\n with"
},
{
"path": "lib/pre_processing.py",
"chars": 3480,
"preview": "###################################################\n#\n# Script to pre-process the original imgs\n#\n####################"
},
{
"path": "lib/pre_processing.py.bak",
"chars": 3479,
"preview": "###################################################\n#\n# Script to pre-process the original imgs\n#\n####################"
},
{
"path": "prepare_datasets_DRIVE.py",
"chars": 4007,
"preview": "#==========================================================\n#\n# This prepare the hdf5 datasets of the DRIVE database\n#\n"
},
{
"path": "src/LadderNetv65.py",
"chars": 7763,
"preview": "import torch\nimport torch.nn.functional as F\nimport torch.nn as nn\n\ndrop = 0.25\n\ndef conv3x3(in_planes, out_planes, stri"
},
{
"path": "src/__init__.py",
"chars": 0,
"preview": ""
},
{
"path": "src/losses.py",
"chars": 1155,
"preview": "import torch\nimport numpy as np\nimport torch.nn as nn\n\ndef cuda(x):\n return x.cuda(async=True) if torch.cuda.is_avail"
},
{
"path": "src/retinaNN_predict.py",
"chars": 12336,
"preview": "###################################################\n#\n# Script to\n# - Calculate prediction of the test dataset\n# -"
},
{
"path": "src/retinaNN_training.py",
"chars": 6531,
"preview": "###################################################\n#\n# Script to:\n# - Load the images and extract the patches\n# -"
},
{
"path": "src/subpixel_upscaling.py",
"chars": 3482,
"preview": "# -*- coding: utf-8 -*-\n\"\"\"\nThis file contains an implementation of Sub-pixel convolutional upscaling layer based on\nthe"
},
{
"path": "test/drive_performances.txt",
"chars": 363,
"preview": "Area under the ROC curve: 0.9793486169167707\nArea under Precision-Recall curve: 0.9107415769226219\nJaccard similarity sc"
},
{
"path": "test/performances.txt",
"chars": 364,
"preview": "Area under the ROC curve: 0.9793257409827817\nArea under Precision-Recall curve: 0.9107420271043044\nJaccard similarity sc"
},
{
"path": "test/test_architecture.json",
"chars": 37810,
"preview": "{\"class_name\": \"Model\", \"config\": {\"name\": \"model_1\", \"layers\": [{\"name\": \"input_1\", \"class_name\": \"InputLayer\", \"config"
},
{
"path": "test/test_configuration.txt",
"chars": 1342,
"preview": "[data paths]\npath_local = ../DRIVE_datasets_training_testing/\ntrain_imgs_original = DRIVE_dataset_imgs_train.hdf5\ntrain"
},
{
"path": "test/test_training.nohup",
"chars": 0,
"preview": ""
}
]
// ... and 43 more files (download for full content)
About this extraction
This page contains the full source code of the juntang-zhuang/LadderNet GitHub repository, extracted and formatted as plain text for AI agents and large language models (LLMs). The extraction includes 75 files (16.3 MB), approximately 58.4k tokens, and a symbol index with 84 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.