├── .idea
    ├── face_swap.iml
    ├── misc.xml
    ├── modules.xml
    └── workspace.xml
├── README.md
├── dataset
    ├── __pycache__
    │   └── reader.cpython-37.pyc
    ├── reader.py
    ├── video_clips_path.txt
    └── video_landmarks_path.txt
├── face-alignment_projection
    ├── conda
    │   ├── conda_upload.sh
    │   └── meta.yaml
    ├── docs
    │   └── images
    │   │   ├── 2dlandmarks.png
    │   │   └── face-alignment-adrian.gif
    ├── examples
    │   └── detect_landmarks_in_image.py
    ├── face_alignment
    │   ├── __init__.py
    │   ├── __pycache__
    │   │   ├── __init__.cpython-36.pyc
    │   │   ├── api.cpython-36.pyc
    │   │   ├── models.cpython-36.pyc
    │   │   └── utils.cpython-36.pyc
    │   ├── api.py
    │   ├── detection
    │   │   ├── __init__.py
    │   │   ├── __pycache__
    │   │   │   ├── __init__.cpython-36.pyc
    │   │   │   └── core.cpython-36.pyc
    │   │   ├── core.py
    │   │   ├── dlib
    │   │   │   ├── __init__.py
    │   │   │   └── dlib_detector.py
    │   │   ├── folder
    │   │   │   ├── __init__.py
    │   │   │   └── folder_detector.py
    │   │   └── sfd
    │   │   │   ├── __init__.py
    │   │   │   ├── __pycache__
    │   │   │       ├── __init__.cpython-36.pyc
    │   │   │       ├── bbox.cpython-36.pyc
    │   │   │       ├── detect.cpython-36.pyc
    │   │   │       ├── net_s3fd.cpython-36.pyc
    │   │   │       └── sfd_detector.cpython-36.pyc
    │   │   │   ├── bbox.py
    │   │   │   ├── detect.py
    │   │   │   ├── net_s3fd.py
    │   │   │   └── sfd_detector.py
    │   ├── models.py
    │   └── utils.py
    ├── get_landmarks.py
    ├── songyiren.jpg
    └── test
    │   ├── assets
    │       └── aflw-test.jpg
    │   ├── facealignment_test.py
    │   ├── smoke_test.py
    │   └── test_utils.py
├── model
    ├── D.py
    ├── E.py
    ├── FaceSwapModel.py
    ├── G.py
    ├── __init__.py
    ├── __pycache__
    │   ├── D.cpython-36.pyc
    │   ├── D.cpython-37.pyc
    │   ├── E.cpython-36.pyc
    │   ├── E.cpython-37.pyc
    │   ├── FaceSwapModel.cpython-37.pyc
    │   ├── G.cpython-36.pyc
    │   ├── G.cpython-37.pyc
    │   ├── __init__.cpython-37.pyc
    │   ├── loss.cpython-36.pyc
    │   ├── loss.cpython-37.pyc
    │   ├── resblocks.cpython-36.pyc
    │   ├── resblocks.cpython-37.pyc
    │   ├── vgg_face.cpython-36.pyc
    │   └── vgg_face.cpython-37.pyc
    ├── loss.py
    ├── resblocks.py
    └── vgg_face.py
├── train.py
├── training_visual
    ├── temp_fake_gt_landmark_100.jpg
    ├── temp_fake_gt_landmark_1000.jpg
    ├── temp_fake_gt_landmark_1100.jpg
    ├── temp_fake_gt_landmark_1200.jpg
    ├── temp_fake_gt_landmark_1300.jpg
    ├── temp_fake_gt_landmark_1400.jpg
    ├── temp_fake_gt_landmark_1500.jpg
    ├── temp_fake_gt_landmark_1600.jpg
    ├── temp_fake_gt_landmark_1700.jpg
    ├── temp_fake_gt_landmark_1800.jpg
    ├── temp_fake_gt_landmark_1900.jpg
    ├── temp_fake_gt_landmark_200.jpg
    ├── temp_fake_gt_landmark_2000.jpg
    ├── temp_fake_gt_landmark_2100.jpg
    ├── temp_fake_gt_landmark_2200.jpg
    ├── temp_fake_gt_landmark_2300.jpg
    ├── temp_fake_gt_landmark_300.jpg
    ├── temp_fake_gt_landmark_400.jpg
    ├── temp_fake_gt_landmark_500.jpg
    ├── temp_fake_gt_landmark_600.jpg
    ├── temp_fake_gt_landmark_700.jpg
    ├── temp_fake_gt_landmark_800.jpg
    └── temp_fake_gt_landmark_900.jpg
└── utils.py


/.idea/face_swap.iml:
--------------------------------------------------------------------------------
 1 | <?xml version="1.0" encoding="UTF-8"?>
 2 | <module type="PYTHON_MODULE" version="4">
 3 |   <component name="NewModuleRootManager">
 4 |     <content url="file://$MODULE_DIR$" />
 5 |     <orderEntry type="jdk" jdkName="Python 3.6" jdkType="Python SDK" />
 6 |     <orderEntry type="sourceFolder" forTests="false" />
 7 |   </component>
 8 |   <component name="TestRunnerService">
 9 |     <option name="projectConfiguration" value="py.test" />
10 |     <option name="PROJECT_TEST_RUNNER" value="py.test" />
11 |   </component>
12 | </module>


--------------------------------------------------------------------------------
/.idea/misc.xml:
--------------------------------------------------------------------------------
1 | <?xml version="1.0" encoding="UTF-8"?>
2 | <project version="4">
3 |   <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.6" project-jdk-type="Python SDK" />
4 | </project>


--------------------------------------------------------------------------------
/.idea/modules.xml:
--------------------------------------------------------------------------------
1 | <?xml version="1.0" encoding="UTF-8"?>
2 | <project version="4">
3 |   <component name="ProjectModuleManager">
4 |     <modules>
5 |       <module fileurl="file://$PROJECT_DIR$/.idea/face_swap.iml" filepath="$PROJECT_DIR$/.idea/face_swap.iml" />
6 |     </modules>
7 |   </component>
8 | </project>


--------------------------------------------------------------------------------
/.idea/workspace.xml:
--------------------------------------------------------------------------------
  1 | <?xml version="1.0" encoding="UTF-8"?>
  2 | <project version="4">
  3 |   <component name="ChangeListManager">
  4 |     <list default="true" id="a5f39a9b-74cc-4546-ac4e-2bcac315d1c0" name="Default" comment="" />
  5 |     <option name="EXCLUDED_CONVERTED_TO_IGNORED" value="true" />
  6 |     <option name="TRACKING_ENABLED" value="true" />
  7 |     <option name="SHOW_DIALOG" value="false" />
  8 |     <option name="HIGHLIGHT_CONFLICTS" value="true" />
  9 |     <option name="HIGHLIGHT_NON_ACTIVE_CHANGELIST" value="false" />
 10 |     <option name="LAST_RESOLUTION" value="IGNORE" />
 11 |   </component>
 12 |   <component name="FileEditorManager">
 13 |     <leaf SIDE_TABS_SIZE_LIMIT_KEY="300">
 14 |       <file leaf-file-name="E.py" pinned="false" current-in-tab="false">
 15 |         <entry file="file://$PROJECT_DIR$/model/E.py">
 16 |           <provider selected="true" editor-type-id="text-editor">
 17 |             <state relative-caret-position="1175">
 18 |               <caret line="54" column="48" selection-start-line="54" selection-start-column="48" selection-end-line="54" selection-end-column="48" />
 19 |               <folding>
 20 |                 <element signature="e#0#12#0" expanded="true" />
 21 |               </folding>
 22 |             </state>
 23 |           </provider>
 24 |         </entry>
 25 |       </file>
 26 |       <file leaf-file-name="train.py" pinned="false" current-in-tab="false">
 27 |         <entry file="file://$PROJECT_DIR$/train.py">
 28 |           <provider selected="true" editor-type-id="text-editor">
 29 |             <state relative-caret-position="166">
 30 |               <caret line="81" selection-start-line="81" selection-end-line="81" />
 31 |               <folding>
 32 |                 <element signature="e#0#12#0" expanded="true" />
 33 |               </folding>
 34 |             </state>
 35 |           </provider>
 36 |         </entry>
 37 |       </file>
 38 |       <file leaf-file-name="loss.py" pinned="false" current-in-tab="false">
 39 |         <entry file="file://$PROJECT_DIR$/model/loss.py">
 40 |           <provider selected="true" editor-type-id="text-editor">
 41 |             <state relative-caret-position="391">
 42 |               <caret line="32" lean-forward="true" selection-start-line="32" selection-end-line="32" />
 43 |             </state>
 44 |           </provider>
 45 |         </entry>
 46 |       </file>
 47 |       <file leaf-file-name="get_landmarks.py" pinned="false" current-in-tab="true">
 48 |         <entry file="file://$PROJECT_DIR$/face-alignment_projection/get_landmarks.py">
 49 |           <provider selected="true" editor-type-id="text-editor">
 50 |             <state relative-caret-position="300">
 51 |               <caret line="17" column="32" lean-forward="true" selection-start-line="17" selection-start-column="32" selection-end-line="17" selection-end-column="32" />
 52 |             </state>
 53 |           </provider>
 54 |         </entry>
 55 |       </file>
 56 |       <file leaf-file-name="G.py" pinned="false" current-in-tab="false">
 57 |         <entry file="file://$PROJECT_DIR$/model/G.py">
 58 |           <provider selected="true" editor-type-id="text-editor">
 59 |             <state relative-caret-position="216">
 60 |               <caret line="72" column="34" lean-forward="true" selection-start-line="72" selection-start-column="34" selection-end-line="72" selection-end-column="34" />
 61 |               <folding>
 62 |                 <element signature="e#0#12#0" expanded="true" />
 63 |               </folding>
 64 |             </state>
 65 |           </provider>
 66 |         </entry>
 67 |       </file>
 68 |       <file leaf-file-name="utils.py" pinned="false" current-in-tab="false">
 69 |         <entry file="file://$PROJECT_DIR$/utils.py">
 70 |           <provider selected="true" editor-type-id="text-editor">
 71 |             <state relative-caret-position="75">
 72 |               <caret line="4" column="42" selection-start-line="4" selection-start-column="42" selection-end-line="4" selection-end-column="42" />
 73 |               <folding>
 74 |                 <element signature="e#0#28#0" expanded="true" />
 75 |               </folding>
 76 |             </state>
 77 |           </provider>
 78 |         </entry>
 79 |       </file>
 80 |       <file leaf-file-name="FaceSwapModel.py" pinned="false" current-in-tab="false">
 81 |         <entry file="file://$PROJECT_DIR$/model/FaceSwapModel.py">
 82 |           <provider selected="true" editor-type-id="text-editor">
 83 |             <state relative-caret-position="1375">
 84 |               <caret line="62" column="25" selection-start-line="62" selection-start-column="25" selection-end-line="62" selection-end-column="25" />
 85 |               <folding>
 86 |                 <element signature="e#0#21#0" expanded="true" />
 87 |               </folding>
 88 |             </state>
 89 |           </provider>
 90 |         </entry>
 91 |       </file>
 92 |       <file leaf-file-name="D.py" pinned="false" current-in-tab="false">
 93 |         <entry file="file://$PROJECT_DIR$/model/D.py">
 94 |           <provider selected="true" editor-type-id="text-editor">
 95 |             <state relative-caret-position="1250">
 96 |               <caret line="57" selection-start-line="57" selection-end-line="57" />
 97 |               <folding>
 98 |                 <element signature="e#0#12#0" expanded="true" />
 99 |               </folding>
100 |             </state>
101 |           </provider>
102 |         </entry>
103 |       </file>
104 |     </leaf>
105 |   </component>
106 |   <component name="FileTemplateManagerImpl">
107 |     <option name="RECENT_TEMPLATES">
108 |       <list>
109 |         <option value="Python Script" />
110 |       </list>
111 |     </option>
112 |   </component>
113 |   <component name="FindInProjectRecents">
114 |     <findStrings>
115 |       <find>Conv2d</find>
116 |       <find>pool</find>
117 |       <find>landmarks</find>
118 |       <find>fa</find>
119 |       <find>v</find>
120 |       <find>Adain</find>
121 |       <find>mean_y</find>
122 |       <find>e</find>
123 |       <find>mean_</find>
124 |       <find>RandomHorizontalFlip</find>
125 |       <find>vis</find>
126 |       <find>print</find>
127 |     </findStrings>
128 |   </component>
129 |   <component name="HighlightingSettingsPerFile">
130 |     <setting file="file://$PROJECT_DIR$/model/G.py" root0="SKIP_INSPECTION" />
131 |   </component>
132 |   <component name="IdeDocumentHistory">
133 |     <option name="CHANGED_PATHS">
134 |       <list>
135 |         <option value="$PROJECT_DIR$/G.py" />
136 |         <option value="$PROJECT_DIR$/model/__init__.py" />
137 |         <option value="$PROJECT_DIR$/model/snresnet64.py" />
138 |         <option value="$PROJECT_DIR$/aaa.py" />
139 |         <option value="$PROJECT_DIR$/../old project/FaceID-GAN/get_key_points/get_points.py" />
140 |         <option value="G:/Anaconda3/Lib/site-packages/face_alignment/api.py" />
141 |         <option value="$PROJECT_DIR$/face-alignment_projection/face_alignment/api.py" />
142 |         <option value="$PROJECT_DIR$/face-alignment_projection/train.py" />
143 |         <option value="$PROJECT_DIR$/dataset/reader.py" />
144 |         <option value="$PROJECT_DIR$/model/E.py" />
145 |         <option value="$PROJECT_DIR$/model/D.py" />
146 |         <option value="$PROJECT_DIR$/model/vgg_face.py" />
147 |         <option value="$PROJECT_DIR$/utils.py" />
148 |         <option value="$PROJECT_DIR$/model/FaceSwapModel.py" />
149 |         <option value="$PROJECT_DIR$/train.py" />
150 |         <option value="$PROJECT_DIR$/model/G.py" />
151 |         <option value="$PROJECT_DIR$/model/loss.py" />
152 |         <option value="$PROJECT_DIR$/face-alignment_projection/get_landmarks.py" />
153 |       </list>
154 |     </option>
155 |   </component>
156 |   <component name="ProjectFrameBounds" extendedState="7">
157 |     <option name="x" value="220" />
158 |     <option name="y" value="35" />
159 |     <option name="width" value="980" />
160 |     <option name="height" value="1040" />
161 |   </component>
162 |   <component name="ProjectView">
163 |     <navigator proportions="" version="1">
164 |       <foldersAlwaysOnTop value="true" />
165 |     </navigator>
166 |     <panes>
167 |       <pane id="Scope" />
168 |       <pane id="ProjectPane">
169 |         <subPane>
170 |           <expand>
171 |             <path>
172 |               <item name="face_swap" type="b2602c69:ProjectViewProjectNode" />
173 |               <item name="face_swap" type="462c0819:PsiDirectoryNode" />
174 |             </path>
175 |             <path>
176 |               <item name="face_swap" type="b2602c69:ProjectViewProjectNode" />
177 |               <item name="face_swap" type="462c0819:PsiDirectoryNode" />
178 |               <item name="model" type="462c0819:PsiDirectoryNode" />
179 |             </path>
180 |           </expand>
181 |           <select />
182 |         </subPane>
183 |       </pane>
184 |     </panes>
185 |   </component>
186 |   <component name="PropertiesComponent">
187 |     <property name="last_opened_file_path" value="$PROJECT_DIR$" />
188 |     <property name="settings.editor.selected.configurable" value="com.jetbrains.python.configuration.PyActiveSdkModuleConfigurable" />
189 |   </component>
190 |   <component name="RecentsManager">
191 |     <key name="MoveFile.RECENT_KEYS">
192 |       <recent name="F:\python\face_swap" />
193 |     </key>
194 |   </component>
195 |   <component name="RunDashboard">
196 |     <option name="ruleStates">
197 |       <list>
198 |         <RuleState>
199 |           <option name="name" value="ConfigurationTypeDashboardGroupingRule" />
200 |         </RuleState>
201 |         <RuleState>
202 |           <option name="name" value="StatusDashboardGroupingRule" />
203 |         </RuleState>
204 |       </list>
205 |     </option>
206 |   </component>
207 |   <component name="RunManager" selected="Python.utils">
208 |     <configuration name="D" type="PythonConfigurationType" factoryName="Python" temporary="true">
209 |       <module name="face_swap" />
210 |       <option name="INTERPRETER_OPTIONS" value="" />
211 |       <option name="PARENT_ENVS" value="true" />
212 |       <envs>
213 |         <env name="PYTHONUNBUFFERED" value="1" />
214 |       </envs>
215 |       <option name="SDK_HOME" value="" />
216 |       <option name="WORKING_DIRECTORY" value="$PROJECT_DIR$/model" />
217 |       <option name="IS_MODULE_SDK" value="true" />
218 |       <option name="ADD_CONTENT_ROOTS" value="true" />
219 |       <option name="ADD_SOURCE_ROOTS" value="true" />
220 |       <option name="SCRIPT_NAME" value="$PROJECT_DIR$/model/D.py" />
221 |       <option name="PARAMETERS" value="" />
222 |       <option name="SHOW_COMMAND_LINE" value="false" />
223 |       <option name="EMULATE_TERMINAL" value="false" />
224 |       <option name="MODULE_MODE" value="false" />
225 |     </configuration>
226 |     <configuration name="E" type="PythonConfigurationType" factoryName="Python" temporary="true">
227 |       <module name="face_swap" />
228 |       <option name="INTERPRETER_OPTIONS" value="" />
229 |       <option name="PARENT_ENVS" value="true" />
230 |       <envs>
231 |         <env name="PYTHONUNBUFFERED" value="1" />
232 |       </envs>
233 |       <option name="SDK_HOME" value="" />
234 |       <option name="WORKING_DIRECTORY" value="$PROJECT_DIR$/model" />
235 |       <option name="IS_MODULE_SDK" value="true" />
236 |       <option name="ADD_CONTENT_ROOTS" value="true" />
237 |       <option name="ADD_SOURCE_ROOTS" value="true" />
238 |       <option name="SCRIPT_NAME" value="$PROJECT_DIR$/model/E.py" />
239 |       <option name="PARAMETERS" value="" />
240 |       <option name="SHOW_COMMAND_LINE" value="false" />
241 |       <option name="EMULATE_TERMINAL" value="false" />
242 |       <option name="MODULE_MODE" value="false" />
243 |     </configuration>
244 |     <configuration name="FaceSwapModel" type="PythonConfigurationType" factoryName="Python" temporary="true">
245 |       <module name="face_swap" />
246 |       <option name="INTERPRETER_OPTIONS" value="" />
247 |       <option name="PARENT_ENVS" value="true" />
248 |       <envs>
249 |         <env name="PYTHONUNBUFFERED" value="1" />
250 |       </envs>
251 |       <option name="SDK_HOME" value="" />
252 |       <option name="WORKING_DIRECTORY" value="$PROJECT_DIR$/model" />
253 |       <option name="IS_MODULE_SDK" value="true" />
254 |       <option name="ADD_CONTENT_ROOTS" value="true" />
255 |       <option name="ADD_SOURCE_ROOTS" value="true" />
256 |       <option name="SCRIPT_NAME" value="$PROJECT_DIR$/model/FaceSwapModel.py" />
257 |       <option name="PARAMETERS" value="" />
258 |       <option name="SHOW_COMMAND_LINE" value="false" />
259 |       <option name="EMULATE_TERMINAL" value="false" />
260 |       <option name="MODULE_MODE" value="false" />
261 |     </configuration>
262 |     <configuration name="reader" type="PythonConfigurationType" factoryName="Python" temporary="true">
263 |       <module name="face_swap" />
264 |       <option name="INTERPRETER_OPTIONS" value="" />
265 |       <option name="PARENT_ENVS" value="true" />
266 |       <envs>
267 |         <env name="PYTHONUNBUFFERED" value="1" />
268 |       </envs>
269 |       <option name="SDK_HOME" value="" />
270 |       <option name="WORKING_DIRECTORY" value="$PROJECT_DIR$/dataset" />
271 |       <option name="IS_MODULE_SDK" value="true" />
272 |       <option name="ADD_CONTENT_ROOTS" value="true" />
273 |       <option name="ADD_SOURCE_ROOTS" value="true" />
274 |       <option name="SCRIPT_NAME" value="$PROJECT_DIR$/dataset/reader.py" />
275 |       <option name="PARAMETERS" value="" />
276 |       <option name="SHOW_COMMAND_LINE" value="false" />
277 |       <option name="EMULATE_TERMINAL" value="false" />
278 |       <option name="MODULE_MODE" value="false" />
279 |     </configuration>
280 |     <configuration name="utils" type="PythonConfigurationType" factoryName="Python" temporary="true">
281 |       <module name="face_swap" />
282 |       <option name="INTERPRETER_OPTIONS" value="" />
283 |       <option name="PARENT_ENVS" value="true" />
284 |       <envs>
285 |         <env name="PYTHONUNBUFFERED" value="1" />
286 |       </envs>
287 |       <option name="SDK_HOME" value="" />
288 |       <option name="WORKING_DIRECTORY" value="$PROJECT_DIR$" />
289 |       <option name="IS_MODULE_SDK" value="true" />
290 |       <option name="ADD_CONTENT_ROOTS" value="true" />
291 |       <option name="ADD_SOURCE_ROOTS" value="true" />
292 |       <option name="SCRIPT_NAME" value="$PROJECT_DIR$/utils.py" />
293 |       <option name="PARAMETERS" value="" />
294 |       <option name="SHOW_COMMAND_LINE" value="false" />
295 |       <option name="EMULATE_TERMINAL" value="false" />
296 |       <option name="MODULE_MODE" value="false" />
297 |     </configuration>
298 |     <list>
299 |       <item itemvalue="Python.E" />
300 |       <item itemvalue="Python.FaceSwapModel" />
301 |       <item itemvalue="Python.reader" />
302 |       <item itemvalue="Python.D" />
303 |       <item itemvalue="Python.utils" />
304 |     </list>
305 |     <recent_temporary>
306 |       <list>
307 |         <item itemvalue="Python.utils" />
308 |         <item itemvalue="Python.FaceSwapModel" />
309 |         <item itemvalue="Python.D" />
310 |         <item itemvalue="Python.E" />
311 |         <item itemvalue="Python.reader" />
312 |       </list>
313 |     </recent_temporary>
314 |   </component>
315 |   <component name="SvnConfiguration">
316 |     <configuration />
317 |   </component>
318 |   <component name="TaskManager">
319 |     <task active="true" id="Default" summary="Default task">
320 |       <changelist id="a5f39a9b-74cc-4546-ac4e-2bcac315d1c0" name="Default" comment="" />
321 |       <created>1559740913359</created>
322 |       <option name="number" value="Default" />
323 |       <option name="presentableId" value="Default" />
324 |       <updated>1559740913359</updated>
325 |     </task>
326 |     <servers />
327 |   </component>
328 |   <component name="ToolWindowManager">
329 |     <frame x="-8" y="-8" width="1936" height="1056" extended-state="6" />
330 |     <editor active="true" />
331 |     <layout>
332 |       <window_info active="true" content_ui="combo" id="Project" order="0" visible="true" weight="0.15104167" />
333 |       <window_info anchor="bottom" id="TODO" order="6" />
334 |       <window_info anchor="bottom" id="Event Log" order="7" side_tool="true" />
335 |       <window_info anchor="bottom" id="Run" order="2" weight="0.32413793" />
336 |       <window_info anchor="bottom" id="Version Control" order="7" show_stripe_button="false" />
337 |       <window_info anchor="bottom" id="Python Console" order="7" />
338 |       <window_info id="Structure" order="1" side_tool="true" weight="0.25" />
339 |       <window_info anchor="bottom" id="Terminal" order="7" />
340 |       <window_info anchor="bottom" id="Debug" order="3" weight="0.3986207" />
341 |       <window_info id="Favorites" order="2" side_tool="true" />
342 |       <window_info anchor="right" content_ui="combo" id="Hierarchy" order="2" weight="0.25" />
343 |       <window_info anchor="right" id="Commander" internal_type="SLIDING" order="0" type="SLIDING" weight="0.4" />
344 |       <window_info anchor="right" id="Ant Build" order="1" weight="0.25" />
345 |       <window_info anchor="bottom" id="Message" order="0" />
346 |       <window_info anchor="bottom" id="Inspection" order="5" weight="0.4" />
347 |       <window_info anchor="bottom" id="Cvs" order="4" weight="0.25" />
348 |       <window_info anchor="bottom" id="Find" order="1" />
349 |     </layout>
350 |     <layout-to-restore>
351 |       <window_info anchor="right" content_ui="combo" id="Hierarchy" order="2" weight="0.25" />
352 |       <window_info anchor="bottom" id="Inspection" order="5" weight="0.4" />
353 |       <window_info active="true" content_ui="combo" id="Project" order="0" visible="true" weight="0.15104167" />
354 |       <window_info id="Structure" order="1" side_tool="true" weight="0.25" />
355 |       <window_info anchor="right" id="Commander" internal_type="SLIDING" order="0" type="SLIDING" weight="0.4" />
356 |       <window_info anchor="right" id="Ant Build" order="1" weight="0.25" />
357 |       <window_info anchor="bottom" id="TODO" order="6" />
358 |       <window_info anchor="bottom" id="Version Control" order="10" show_stripe_button="false" />
359 |       <window_info anchor="bottom" id="Run" order="2" visible="true" weight="0.32413793" />
360 |       <window_info anchor="bottom" id="Message" order="0" />
361 |       <window_info anchor="bottom" id="Debug" order="3" weight="0.3995749" />
362 |       <window_info anchor="bottom" id="Terminal" order="8" />
363 |       <window_info anchor="bottom" id="Event Log" order="9" side_tool="true" />
364 |       <window_info anchor="bottom" id="Cvs" order="4" weight="0.25" />
365 |       <window_info id="Favorites" order="2" side_tool="true" />
366 |       <window_info anchor="bottom" id="Find" order="1" />
367 |       <window_info anchor="bottom" id="Python Console" order="7" />
368 |     </layout-to-restore>
369 |   </component>
370 |   <component name="VcsContentAnnotationSettings">
371 |     <option name="myLimit" value="2678400000" />
372 |   </component>
373 |   <component name="XDebuggerManager">
374 |     <breakpoint-manager>
375 |       <option name="time" value="5" />
376 |     </breakpoint-manager>
377 |   </component>
378 |   <component name="editorHistoryManager">
379 |     <entry file="file://$PROJECT_DIR$/../pytorch-AdaIN/function.py">
380 |       <provider selected="true" editor-type-id="text-editor">
381 |         <state relative-caret-position="69">
382 |           <caret line="3" selection-start-line="3" selection-end-line="11" selection-end-column="30" />
383 |         </state>
384 |       </provider>
385 |     </entry>
386 |     <entry file="file://$PROJECT_DIR$/model/snresnet64.py" />
387 |     <entry file="file://G:/Anaconda3/Lib/site-packages/torch/nn/parameter.py">
388 |       <provider selected="true" editor-type-id="text-editor">
389 |         <state relative-caret-position="45">
390 |           <caret line="4" column="6" selection-start-line="4" selection-start-column="6" selection-end-line="4" selection-end-column="6" />
391 |         </state>
392 |       </provider>
393 |     </entry>
394 |     <entry file="file://$PROJECT_DIR$/../Face_Recognition/ArcFace/mobile_model.py">
395 |       <provider selected="true" editor-type-id="text-editor">
396 |         <state relative-caret-position="345">
397 |           <caret line="210" column="23" selection-start-line="210" selection-start-column="23" selection-end-line="210" selection-end-column="23" />
398 |         </state>
399 |       </provider>
400 |     </entry>
401 |     <entry file="file://G:/Anaconda3/Lib/site-packages/torch/tensor.py">
402 |       <provider selected="true" editor-type-id="text-editor">
403 |         <state relative-caret-position="-759">
404 |           <caret line="14" column="6" selection-start-line="14" selection-start-column="6" selection-end-line="14" selection-end-column="6" />
405 |         </state>
406 |       </provider>
407 |     </entry>
408 |     <entry file="file://$PROJECT_DIR$/../cgan_with_projection/models/discriminators/snresnet64.py">
409 |       <provider selected="true" editor-type-id="text-editor">
410 |         <state relative-caret-position="184">
411 |           <caret line="53" column="21" selection-start-line="53" selection-start-column="15" selection-end-line="53" selection-end-column="21" />
412 |         </state>
413 |       </provider>
414 |     </entry>
415 |     <entry file="file://$PROJECT_DIR$/../cgan_with_projection/train.py">
416 |       <provider selected="true" editor-type-id="text-editor">
417 |         <state relative-caret-position="184">
418 |           <caret line="143" column="32" selection-start-line="143" selection-start-column="29" selection-end-line="143" selection-end-column="32" />
419 |         </state>
420 |       </provider>
421 |     </entry>
422 |     <entry file="file://G:/Anaconda3/Lib/site-packages/torch/nn/modules/loss.py">
423 |       <provider selected="true" editor-type-id="text-editor">
424 |         <state relative-caret-position="99">
425 |           <caret line="87" column="8" selection-start-line="87" selection-start-column="8" selection-end-line="87" selection-end-column="8" />
426 |         </state>
427 |       </provider>
428 |     </entry>
429 |     <entry file="file://$PROJECT_DIR$/aaa.py" />
430 |     <entry file="file://G:/Anaconda3/Lib/site-packages/torch/nn/modules/conv.py">
431 |       <provider selected="true" editor-type-id="text-editor">
432 |         <state relative-caret-position="214">
433 |           <caret line="319" selection-start-line="319" selection-end-line="319" />
434 |         </state>
435 |       </provider>
436 |     </entry>
437 |     <entry file="file://G:/Anaconda3/Lib/site-packages/torch/utils/model_zoo.py">
438 |       <provider selected="true" editor-type-id="text-editor">
439 |         <state relative-caret-position="-590" />
440 |       </provider>
441 |     </entry>
442 |     <entry file="file://$PROJECT_DIR$/../DANet-master/encoding/datasets/pascal_aug.py">
443 |       <provider selected="true" editor-type-id="text-editor">
444 |         <state relative-caret-position="-1059" />
445 |       </provider>
446 |     </entry>
447 |     <entry file="file://$PROJECT_DIR$/../DANet-master/danet/train.py">
448 |       <provider selected="true" editor-type-id="text-editor">
449 |         <state relative-caret-position="450">
450 |           <caret line="39" column="43" selection-start-line="39" selection-start-column="43" selection-end-line="39" selection-end-column="43" />
451 |         </state>
452 |       </provider>
453 |     </entry>
454 |     <entry file="file://$PROJECT_DIR$/face-alignment/setup.py" />
455 |     <entry file="file://$PROJECT_DIR$/face-alignment_projection/songyiren.jpg">
456 |       <provider selected="true" editor-type-id="images" />
457 |     </entry>
458 |     <entry file="file://$PROJECT_DIR$/face-alignment_projection/test/test_utils.py">
459 |       <provider selected="true" editor-type-id="text-editor" />
460 |     </entry>
461 |     <entry file="file://G:/Anaconda3/Lib/site-packages/face_alignment/models.py">
462 |       <provider selected="true" editor-type-id="text-editor">
463 |         <state relative-caret-position="150">
464 |           <caret line="154" column="40" lean-forward="true" selection-start-line="154" selection-start-column="40" selection-end-line="154" selection-end-column="40" />
465 |         </state>
466 |       </provider>
467 |     </entry>
468 |     <entry file="file://$PROJECT_DIR$/face-alignment_projection/examples/detect_landmarks_in_image.py">
469 |       <provider selected="true" editor-type-id="text-editor">
470 |         <state relative-caret-position="375">
471 |           <caret line="15" column="16" lean-forward="true" selection-start-line="15" selection-start-column="16" selection-end-line="15" selection-end-column="16" />
472 |         </state>
473 |       </provider>
474 |     </entry>
475 |     <entry file="file://G:/Anaconda3/Lib/os.py">
476 |       <provider selected="true" editor-type-id="text-editor">
477 |         <state relative-caret-position="25">
478 |           <caret line="753" selection-start-line="753" selection-end-line="753" />
479 |         </state>
480 |       </provider>
481 |     </entry>
482 |     <entry file="file://G:/Anaconda3/Lib/site-packages/face_alignment/utils.py">
483 |       <provider selected="true" editor-type-id="text-editor">
484 |         <state relative-caret-position="25">
485 |           <caret line="236" column="25" selection-start-line="236" selection-start-column="25" selection-end-line="236" selection-end-column="25" />
486 |         </state>
487 |       </provider>
488 |     </entry>
489 |     <entry file="file://G:/Anaconda3/Lib/site-packages/numpy/lib/_datasource.py">
490 |       <provider selected="true" editor-type-id="text-editor">
491 |         <state relative-caret-position="-667">
492 |           <caret line="451" column="8" selection-start-line="451" selection-start-column="8" selection-end-line="451" selection-end-column="8" />
493 |         </state>
494 |       </provider>
495 |     </entry>
496 |     <entry file="file://G:/Anaconda3/Lib/site-packages/skimage/io/_io.py">
497 |       <provider selected="true" editor-type-id="text-editor">
498 |         <state relative-caret-position="400">
499 |           <caret line="67" column="15" selection-start-line="67" selection-start-column="12" selection-end-line="67" selection-end-column="15" />
500 |         </state>
501 |       </provider>
502 |     </entry>
503 |     <entry file="file://G:/Anaconda3/Lib/site-packages/skimage/io/manage_plugins.py">
504 |       <provider selected="true" editor-type-id="text-editor">
505 |         <state relative-caret-position="25">
506 |           <caret line="174" column="4" selection-start-line="174" selection-start-column="4" selection-end-line="174" selection-end-column="4" />
507 |         </state>
508 |       </provider>
509 |     </entry>
510 |     <entry file="file://$PROJECT_DIR$/face-alignment_projection/test/facealignment_test.py">
511 |       <provider selected="true" editor-type-id="text-editor">
512 |         <state relative-caret-position="150">
513 |           <caret line="7" column="24" selection-start-line="7" selection-start-column="11" selection-end-line="7" selection-end-column="24" />
514 |         </state>
515 |       </provider>
516 |     </entry>
517 |     <entry file="file://$PROJECT_DIR$/../old project/FaceID-GAN/get_key_points/get_points.py">
518 |       <provider selected="true" editor-type-id="text-editor">
519 |         <state relative-caret-position="225">
520 |           <caret line="27" column="9" selection-start-line="27" selection-start-column="4" selection-end-line="27" selection-end-column="9" />
521 |         </state>
522 |       </provider>
523 |     </entry>
524 |     <entry file="file://G:/Anaconda3/Lib/site-packages/face_alignment/api.py">
525 |       <provider selected="true" editor-type-id="text-editor">
526 |         <state relative-caret-position="225">
527 |           <caret line="178" column="46" selection-start-line="178" selection-start-column="46" selection-end-line="178" selection-end-column="46" />
528 |         </state>
529 |       </provider>
530 |     </entry>
531 |     <entry file="file://$PROJECT_DIR$/face-alignment_projection/face_alignment/api.py">
532 |       <provider selected="true" editor-type-id="text-editor">
533 |         <state relative-caret-position="165">
534 |           <caret line="130" column="27" lean-forward="true" selection-start-line="130" selection-start-column="27" selection-end-line="130" selection-end-column="27" />
535 |         </state>
536 |       </provider>
537 |     </entry>
538 |     <entry file="file://$PROJECT_DIR$/face-alignment_projection/face_alignment/detection/core.py">
539 |       <provider selected="true" editor-type-id="text-editor">
540 |         <state relative-caret-position="25">
541 |           <caret line="32" column="8" selection-start-line="32" selection-start-column="8" selection-end-line="32" selection-end-column="8" />
542 |         </state>
543 |       </provider>
544 |     </entry>
545 |     <entry file="file://$PROJECT_DIR$/face-alignment_projection/face_alignment/detection/dlib/dlib_detector.py">
546 |       <provider selected="true" editor-type-id="text-editor">
547 |         <state relative-caret-position="71">
548 |           <caret line="45" column="8" selection-start-line="45" selection-start-column="8" selection-end-line="45" selection-end-column="8" />
549 |         </state>
550 |       </provider>
551 |     </entry>
552 |     <entry file="file://$PROJECT_DIR$/model/resblocks.py">
553 |       <provider selected="true" editor-type-id="text-editor">
554 |         <state relative-caret-position="475">
555 |           <caret line="73" column="28" selection-start-line="73" selection-start-column="20" selection-end-line="73" selection-end-column="28" />
556 |         </state>
557 |       </provider>
558 |     </entry>
559 |     <entry file="file://$PROJECT_DIR$/face-alignment_projection/test/smoke_test.py">
560 |       <provider selected="true" editor-type-id="text-editor" />
561 |     </entry>
562 |     <entry file="file://G:/Anaconda3/Lib/site-packages/torch/nn/modules/container.py">
563 |       <provider selected="true" editor-type-id="text-editor">
564 |         <state relative-caret-position="-596">
565 |           <caret line="54" column="24" selection-start-line="54" selection-start-column="8" selection-end-line="54" selection-end-column="24" />
566 |         </state>
567 |       </provider>
568 |     </entry>
569 |     <entry file="file://G:/Anaconda3/Lib/site-packages/torchvision/transforms/transforms.py">
570 |       <provider selected="true" editor-type-id="text-editor">
571 |         <state relative-caret-position="212">
572 |           <caret line="43" column="8" selection-start-line="43" selection-start-column="8" selection-end-line="43" selection-end-column="8" />
573 |         </state>
574 |       </provider>
575 |     </entry>
576 |     <entry file="file://G:/Anaconda3/Lib/site-packages/torchvision/transforms/functional.py">
577 |       <provider selected="true" editor-type-id="text-editor">
578 |         <state relative-caret-position="25">
579 |           <caret line="4" column="45" lean-forward="true" selection-start-line="4" selection-start-column="45" selection-end-line="4" selection-end-column="45" />
580 |         </state>
581 |       </provider>
582 |     </entry>
583 |     <entry file="file://$PROJECT_DIR$/dataset/reader.py">
584 |       <provider selected="true" editor-type-id="text-editor">
585 |         <state relative-caret-position="600">
586 |           <caret line="76" column="23" selection-start-line="76" selection-start-column="13" selection-end-line="76" selection-end-column="23" />
587 |         </state>
588 |       </provider>
589 |     </entry>
590 |     <entry file="file://G:/Anaconda3/Lib/site-packages/torch/nn/modules/module.py">
591 |       <provider selected="true" editor-type-id="text-editor">
592 |         <state relative-caret-position="-50">
593 |           <caret line="254" column="36" selection-start-line="254" selection-start-column="36" selection-end-line="254" selection-end-column="36" />
594 |         </state>
595 |       </provider>
596 |     </entry>
597 |     <entry file="file://G:/Anaconda3/Lib/site-packages/torchvision/models/__init__.py">
598 |       <provider selected="true" editor-type-id="text-editor">
599 |         <state relative-caret-position="50">
600 |           <caret line="2" column="9" selection-start-line="2" selection-start-column="6" selection-end-line="2" selection-end-column="9" />
601 |         </state>
602 |       </provider>
603 |     </entry>
604 |     <entry file="file://G:/Anaconda3/Lib/site-packages/torchvision/models/vgg.py">
605 |       <provider selected="true" editor-type-id="text-editor">
606 |         <state relative-caret-position="70">
607 |           <caret line="18" column="29" selection-start-line="18" selection-start-column="29" selection-end-line="18" selection-end-column="29" />
608 |         </state>
609 |       </provider>
610 |     </entry>
611 |     <entry file="file://$PROJECT_DIR$/../old project/dr-gan/train.py">
612 |       <provider selected="true" editor-type-id="text-editor">
613 |         <state relative-caret-position="287">
614 |           <caret line="94" column="36" selection-start-line="94" selection-start-column="27" selection-end-line="94" selection-end-column="36" />
615 |         </state>
616 |       </provider>
617 |     </entry>
618 |     <entry file="file://$PROJECT_DIR$/model/vgg_face.py">
619 |       <provider selected="true" editor-type-id="text-editor">
620 |         <state relative-caret-position="1075">
621 |           <caret line="44" column="24" selection-start-line="44" selection-start-column="18" selection-end-line="44" selection-end-column="24" />
622 |         </state>
623 |       </provider>
624 |     </entry>
625 |     <entry file="file://$PROJECT_DIR$/utils.py">
626 |       <provider selected="true" editor-type-id="text-editor">
627 |         <state relative-caret-position="75">
628 |           <caret line="4" column="42" selection-start-line="4" selection-start-column="42" selection-end-line="4" selection-end-column="42" />
629 |           <folding>
630 |             <element signature="e#0#28#0" expanded="true" />
631 |           </folding>
632 |         </state>
633 |       </provider>
634 |     </entry>
635 |     <entry file="file://$PROJECT_DIR$/model/D.py">
636 |       <provider selected="true" editor-type-id="text-editor">
637 |         <state relative-caret-position="1250">
638 |           <caret line="57" selection-start-line="57" selection-end-line="57" />
639 |           <folding>
640 |             <element signature="e#0#12#0" expanded="true" />
641 |           </folding>
642 |         </state>
643 |       </provider>
644 |     </entry>
645 |     <entry file="file://G:/Anaconda3/Lib/site-packages/numpy/core/shape_base.py">
646 |       <provider selected="true" editor-type-id="text-editor">
647 |         <state relative-caret-position="-296">
648 |           <caret line="290" column="4" selection-start-line="290" selection-start-column="4" selection-end-line="290" selection-end-column="4" />
649 |         </state>
650 |       </provider>
651 |     </entry>
652 |     <entry file="file://G:/Anaconda3/Lib/site-packages/visdom/__init__.py">
653 |       <provider selected="true" editor-type-id="text-editor">
654 |         <state relative-caret-position="25">
655 |           <caret line="930" column="33" selection-start-line="930" selection-start-column="29" selection-end-line="930" selection-end-column="33" />
656 |         </state>
657 |       </provider>
658 |     </entry>
659 |     <entry file="file://$PROJECT_DIR$/model/E.py">
660 |       <provider selected="true" editor-type-id="text-editor">
661 |         <state relative-caret-position="1175">
662 |           <caret line="54" column="48" selection-start-line="54" selection-start-column="48" selection-end-line="54" selection-end-column="48" />
663 |           <folding>
664 |             <element signature="e#0#12#0" expanded="true" />
665 |           </folding>
666 |         </state>
667 |       </provider>
668 |     </entry>
669 |     <entry file="file://$PROJECT_DIR$/model/FaceSwapModel.py">
670 |       <provider selected="true" editor-type-id="text-editor">
671 |         <state relative-caret-position="1375">
672 |           <caret line="62" column="25" selection-start-line="62" selection-start-column="25" selection-end-line="62" selection-end-column="25" />
673 |           <folding>
674 |             <element signature="e#0#21#0" expanded="true" />
675 |           </folding>
676 |         </state>
677 |       </provider>
678 |     </entry>
679 |     <entry file="file://$PROJECT_DIR$/train.py">
680 |       <provider selected="true" editor-type-id="text-editor">
681 |         <state relative-caret-position="166">
682 |           <caret line="81" selection-start-line="81" selection-end-line="81" />
683 |           <folding>
684 |             <element signature="e#0#12#0" expanded="true" />
685 |           </folding>
686 |         </state>
687 |       </provider>
688 |     </entry>
689 |     <entry file="file://$PROJECT_DIR$/model/G.py">
690 |       <provider selected="true" editor-type-id="text-editor">
691 |         <state relative-caret-position="216">
692 |           <caret line="72" column="34" lean-forward="true" selection-start-line="72" selection-start-column="34" selection-end-line="72" selection-end-column="34" />
693 |           <folding>
694 |             <element signature="e#0#12#0" expanded="true" />
695 |           </folding>
696 |         </state>
697 |       </provider>
698 |     </entry>
699 |     <entry file="file://$PROJECT_DIR$/face-alignment_projection/setup.py">
700 |       <provider selected="true" editor-type-id="text-editor">
701 |         <state relative-caret-position="-1334">
702 |           <folding>
703 |             <element signature="e#0#9#0" expanded="true" />
704 |           </folding>
705 |         </state>
706 |       </provider>
707 |     </entry>
708 |     <entry file="file://$PROJECT_DIR$/model/loss.py">
709 |       <provider selected="true" editor-type-id="text-editor">
710 |         <state relative-caret-position="391">
711 |           <caret line="32" lean-forward="true" selection-start-line="32" selection-end-line="32" />
712 |         </state>
713 |       </provider>
714 |     </entry>
715 |     <entry file="file://$PROJECT_DIR$/face-alignment_projection/get_landmarks.py">
716 |       <provider selected="true" editor-type-id="text-editor">
717 |         <state relative-caret-position="300">
718 |           <caret line="17" column="32" lean-forward="true" selection-start-line="17" selection-start-column="32" selection-end-line="17" selection-end-column="32" />
719 |         </state>
720 |       </provider>
721 |     </entry>
722 |   </component>
723 | </project>


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
 1 | # Few-Shot-Adversarial-Learning-for-face-swap
 2 | This is a unofficial re-implementation of the paper "Few-Shot Adversarial Learning of Realistic Neural Talking Head Models" based on Pytorch.
 3 | 
 4 | # description
 5 | The paper from SAMSUNG AI lab presents a new novel and efficient method for face swap, which has amazing performance. I am a student and interested in this paper, so I try to reproduce it.
 6 | I have writen the script for getting landmarks dataset, and dataloader pipeline. I also construct the whole network including Embedding, Generator and Discriminator with loss functions. Everything is done according to the paper.
 7 | But, due to unspecific network descripted by the paper, I found that there are some mistakes in Generator, especially for the Adaptive instance normalization. The training process gets weird results, which you can look at it from "training_visual" file.
 8 | **For getting more understanding about this amazing and great work, I open source this projection and invite more people who are interested in it to become contributor.** If someone reproduces it successfully in the future, please tell me. Thanks!
 9 | 
10 | # how to use
11 | 
12 | 1. get landmarks information.
13 | 
14 | you should download the [VoxCeleb, about 36GB](http://www.robots.ox.ac.uk/~vgg/research/CMBiometrics/data/dense-face-frames.tar.gz). And after unzip the dataset, please change the name of root path of the dataset as "voxcelb1", or you can adapt "./face-alignment_projection/get_landmarks.py" for your envirnoment.
15 | > python get_landmarks.py
16 | 
17 | For test you shouldn't preprocess all data in the VoxCeleb dataset, because it's time-consuming. I just use about 200 video clips. The two files are generated. "video_clips_path.txt" records all clips which have been preprocessed. "video_landmarks_path.txt" records landmarks information path.
18 | 
19 | 2. download VGGFace weights for perceptual loss.
20 | 
21 | you can download from [here](http://www.robots.ox.ac.uk/~albanie/models/pytorch-mcn/vgg_m_face_bn_dag.py) [on this website](http://www.robots.ox.ac.uk/~albanie/pytorch-models.html).
22 | make sure that put the weight file under the "./pretrained" file path.
23 | 
24 | 3. start to train the whole network.
25 | 
26 | > python ./train.py
27 | 
28 | For each 100 iteartion, you can look at temp results including fake image, GT and landmarks under the "./training_visual" file path.
29 | 
30 | # Cite
31 | [1] Few-Shot Adversarial Learning of Realistic Neural Talking Head Models
32 | 
33 | [2] face-alignment is from [the great work](https://github.com/1adrianb/face-alignment)
34 | 


--------------------------------------------------------------------------------
/dataset/__pycache__/reader.cpython-37.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/dataset/__pycache__/reader.cpython-37.pyc


--------------------------------------------------------------------------------
/dataset/reader.py:
--------------------------------------------------------------------------------
 1 | from torch.utils.data import Dataset,DataLoader
 2 | import torch
 3 | import os
 4 | import cv2
 5 | import numpy as np
 6 | from PIL import Image
 7 | from torchvision.transforms import transforms
 8 | from torchvision.transforms import functional as F
 9 | import random
10 | 
11 | class Reader(Dataset):
12 |     def __init__(self,clip_txt,transform):
13 |         super(Reader,self).__init__()
14 |         f = open(clip_txt, 'r')
15 |         clip_list = f.readlines()
16 |         f.close()
17 |         self.clip_list = {}
18 |         self.landmarks_list = []
19 |         for c in clip_list:
20 |             c = c.strip()
21 |             d = c.split('data')
22 |             d = d[0] + 'dataset/voxcelb1/' 'landmarks' + d[2]
23 |             imgs = os.listdir(c)
24 |             self.clip_list[c] = imgs
25 |             self.landmarks_list.append(d)
26 |         self.trans = transform
27 |         self.flip = F.hflip
28 |     def __getitem__(self, index):
29 |         landmark_path = self.landmarks_list[index]
30 |         clip_path = landmark_path.split('landmarks')
31 |         clip_path = clip_path[0] + 'data' + clip_path[1]
32 |         clip_imgs = self.clip_list[clip_path]
33 |         np.random.shuffle(clip_imgs)
34 |         imgs_for_e_path = clip_imgs[:8]
35 |         imgs_for_training_path = clip_imgs[8:]
36 |         imgs_for_e = []
37 |         landmarks_for_e = []
38 |         imgs_for_training = []
39 |         landmarks_for_training = []
40 |         for p in imgs_for_e_path:
41 |             img = Image.open(os.path.join(clip_path,p)).convert('RGB')
42 |             landmark = Image.open(os.path.join(landmark_path,p)).convert('RGB')
43 |             is_flip = random.random() > 0.5
44 |             if is_flip:
45 |                 img = self.flip(img)
46 |                 landmark = self.flip(landmark)
47 |             img = self.trans(img)
48 |             landmark = self.trans(landmark)
49 |             imgs_for_e.append(img)
50 |             landmarks_for_e.append(landmark)
51 |         for p in imgs_for_training_path:
52 |             img = Image.open(os.path.join(clip_path, p)).convert('RGB')
53 |             landmark = Image.open(os.path.join(landmark_path, p)).convert('RGB')
54 |             is_flip = random.random() > 0.5
55 |             if is_flip:
56 |                 img = self.flip(img)
57 |                 landmark = self.flip(landmark)
58 |             img = self.trans(img)
59 |             landmark = self.trans(landmark)
60 |             imgs_for_training.append(img)
61 |             landmarks_for_training.append(landmark)
62 |         return {'imgs_e':imgs_for_e,
63 |                 'landmarks_e':landmarks_for_e,
64 |                 'imgs_training':imgs_for_training,
65 |                 'landmarks_training':landmarks_for_training
66 |                 }
67 |     def __len__(self):
68 |         return len(self.landmarks_list)
69 | 
70 | def get_loader(clip_txt,batchsize,num_workers):
71 |     trans = transforms.Compose([
72 |         transforms.Resize((256,256)),
73 |         transforms.ToTensor(),
74 |         transforms.Normalize([0.5,0.5,0.5],[0.5,0.5,0.5])
75 |     ])
76 |     dataset = Reader(clip_txt,trans)
77 |     loader = DataLoader(dataset,batchsize,pin_memory=True,drop_last=True,num_workers=num_workers)
78 |     return loader, len(dataset)
79 | if __name__=='__main__':
80 |     clip_txt = 'video_clips_path.txt'
81 |     batchsize = 1
82 |     num_workers = 0
83 |     loader,_ = get_loader(clip_txt,batchsize,num_workers)
84 |     for d in loader:
85 |         for k, v in d.items():
86 |             print(k)
87 |             print(type(v))
88 |             print(type(v[0]))
89 |             print(v[0].size())  # torch.Size([1, 3, 256, 256])
90 | 
91 |             print('**********')
92 | 


--------------------------------------------------------------------------------
/face-alignment_projection/conda/conda_upload.sh:
--------------------------------------------------------------------------------
1 | PCG_NAME=face_alignment
2 | USER=1adrianb
3 | 
4 | mkdir ~/conda-build
5 | conda config --set anaconda_upload no
6 | conda build conda/
7 | anaconda -t $CONDA_UPLOAD_TOKEN upload -u $USER /home/travis/miniconda/envs/test-environment/conda-bld/noarch/face_alignment-1.0.1-py_1.tar.bz2 --force


--------------------------------------------------------------------------------
/face-alignment_projection/conda/meta.yaml:
--------------------------------------------------------------------------------
 1 | {% set version = "1.0.1" %}
 2 | 
 3 | package:
 4 |   name: face_alignment
 5 |   version: {{ version }}
 6 | 
 7 | source:
 8 |   path: ..
 9 | 
10 | build:
11 |   number: 1
12 |   noarch: python
13 |   script: python setup.py install --single-version-externally-managed --record=record.txt
14 | 
15 | requirements:
16 |   build:
17 |     - setuptools
18 |     - python
19 |   run:
20 |     - python
21 |     - pytorch
22 |     - numpy
23 |     - scikit-image
24 |     - scipy
25 |     - opencv
26 |     - tqdm
27 | 
28 | about:
29 |   home: https://github.com/1adrianb/face-alignment
30 |   license: BSD
31 |   license_file: LICENSE
32 |   summary: A 2D and 3D face alignment libray in python
33 | 
34 | extra:
35 |   recipe-maintainers:
36 |     - 1adrianb
37 | 


--------------------------------------------------------------------------------
/face-alignment_projection/docs/images/2dlandmarks.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/docs/images/2dlandmarks.png


--------------------------------------------------------------------------------
/face-alignment_projection/docs/images/face-alignment-adrian.gif:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/docs/images/face-alignment-adrian.gif


--------------------------------------------------------------------------------
/face-alignment_projection/examples/detect_landmarks_in_image.py:
--------------------------------------------------------------------------------
 1 | import face_alignment
 2 | import numpy as np
 3 | from mpl_toolkits.mplot3d import Axes3D
 4 | import matplotlib.pyplot as plt
 5 | from skimage import io
 6 | 
 7 | # Run the 3D face alignment on a test image, without CUDA.
 8 | fa = face_alignment.FaceAlignment(face_alignment.LandmarksType._3D, device='cuda:0', flip_input=True)
 9 | 
10 | input = io.imread('../test/assets/aflw-test.jpg')
11 | preds = fa.get_landmarks(input)[-1]
12 | 
13 | #TODO: Make this nice
14 | fig = plt.figure(figsize=plt.figaspect(.5))
15 | ax = fig.add_subplot(1, 2, 1)
16 | ax.imshow(input)
17 | ax.plot(preds[0:17,0],preds[0:17,1],marker='o',markersize=6,linestyle='-',color='w',lw=2)
18 | ax.plot(preds[17:22,0],preds[17:22,1],marker='o',markersize=6,linestyle='-',color='w',lw=2)
19 | ax.plot(preds[22:27,0],preds[22:27,1],marker='o',markersize=6,linestyle='-',color='w',lw=2)
20 | ax.plot(preds[27:31,0],preds[27:31,1],marker='o',markersize=6,linestyle='-',color='w',lw=2)
21 | ax.plot(preds[31:36,0],preds[31:36,1],marker='o',markersize=6,linestyle='-',color='w',lw=2)
22 | ax.plot(preds[36:42,0],preds[36:42,1],marker='o',markersize=6,linestyle='-',color='w',lw=2)
23 | ax.plot(preds[42:48,0],preds[42:48,1],marker='o',markersize=6,linestyle='-',color='w',lw=2)
24 | ax.plot(preds[48:60,0],preds[48:60,1],marker='o',markersize=6,linestyle='-',color='w',lw=2)
25 | ax.plot(preds[60:68,0],preds[60:68,1],marker='o',markersize=6,linestyle='-',color='w',lw=2) 
26 | ax.axis('off')
27 | 
28 | ax = fig.add_subplot(1, 2, 2, projection='3d')
29 | surf = ax.scatter(preds[:,0]*1.2,preds[:,1],preds[:,2],c="cyan", alpha=1.0, edgecolor='b')
30 | ax.plot3D(preds[:17,0]*1.2,preds[:17,1], preds[:17,2], color='blue' )
31 | ax.plot3D(preds[17:22,0]*1.2,preds[17:22,1],preds[17:22,2], color='blue')
32 | ax.plot3D(preds[22:27,0]*1.2,preds[22:27,1],preds[22:27,2], color='blue')
33 | ax.plot3D(preds[27:31,0]*1.2,preds[27:31,1],preds[27:31,2], color='blue')
34 | ax.plot3D(preds[31:36,0]*1.2,preds[31:36,1],preds[31:36,2], color='blue')
35 | ax.plot3D(preds[36:42,0]*1.2,preds[36:42,1],preds[36:42,2], color='blue')
36 | ax.plot3D(preds[42:48,0]*1.2,preds[42:48,1],preds[42:48,2], color='blue')
37 | ax.plot3D(preds[48:,0]*1.2,preds[48:,1],preds[48:,2], color='blue' )
38 | 
39 | ax.view_init(elev=90., azim=90.)
40 | ax.set_xlim(ax.get_xlim()[::-1])
41 | plt.show()
42 | 


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/__init__.py:
--------------------------------------------------------------------------------
1 | # -*- coding: utf-8 -*-
2 | 
3 | __author__ = """Adrian Bulat"""
4 | __email__ = 'adrian.bulat@nottingham.ac.uk'
5 | __version__ = '1.0.1'
6 | 
7 | from .api import FaceAlignment, LandmarksType, NetworkSize
8 | 


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/__pycache__/__init__.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/face_alignment/__pycache__/__init__.cpython-36.pyc


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/__pycache__/api.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/face_alignment/__pycache__/api.cpython-36.pyc


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/__pycache__/models.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/face_alignment/__pycache__/models.cpython-36.pyc


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/__pycache__/utils.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/face_alignment/__pycache__/utils.cpython-36.pyc


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/api.py:
--------------------------------------------------------------------------------
  1 | from __future__ import print_function
  2 | import os
  3 | import torch
  4 | from torch.utils.model_zoo import load_url
  5 | from enum import Enum
  6 | from skimage import io
  7 | from skimage import color
  8 | import numpy as np
  9 | import cv2
 10 | try:
 11 |     import urllib.request as request_file
 12 | except BaseException:
 13 |     import urllib as request_file
 14 | 
 15 | from .models import FAN, ResNetDepth
 16 | from .utils import *
 17 | 
 18 | 
 19 | class LandmarksType(Enum):
 20 |     """Enum class defining the type of landmarks to detect.
 21 | 
 22 |     ``_2D`` - the detected points ``(x,y)`` are detected in a 2D space and follow the visible contour of the face
 23 |     ``_2halfD`` - this points represent the projection of the 3D points into 3D
 24 |     ``_3D`` - detect the points ``(x,y,z)``` in a 3D space
 25 | 
 26 |     """
 27 |     _2D = 1
 28 |     _2halfD = 2
 29 |     _3D = 3
 30 | 
 31 | 
 32 | class NetworkSize(Enum):
 33 |     # TINY = 1
 34 |     # SMALL = 2
 35 |     # MEDIUM = 3
 36 |     LARGE = 4
 37 | 
 38 |     def __new__(cls, value):
 39 |         member = object.__new__(cls)
 40 |         member._value_ = value
 41 |         return member
 42 | 
 43 |     def __int__(self):
 44 |         return self.value
 45 | 
 46 | models_urls = {
 47 |     '2DFAN-4': 'https://www.adrianbulat.com/downloads/python-fan/2DFAN4-11f355bf06.pth.tar',
 48 |     '3DFAN-4': 'https://www.adrianbulat.com/downloads/python-fan/3DFAN4-7835d9f11d.pth.tar',
 49 |     'depth': 'https://www.adrianbulat.com/downloads/python-fan/depth-2a464da4ea.pth.tar',
 50 | }
 51 | 
 52 | 
 53 | class FaceAlignment:
 54 |     def __init__(self, landmarks_type, network_size=NetworkSize.LARGE,
 55 |                  device='cuda', flip_input=False, face_detector='sfd', verbose=False):
 56 |         self.device = device
 57 |         self.flip_input = flip_input
 58 |         self.landmarks_type = landmarks_type
 59 |         self.verbose = verbose
 60 | 
 61 |         network_size = int(network_size)
 62 | 
 63 |         if 'cuda' in device:
 64 |             torch.backends.cudnn.benchmark = True
 65 | 
 66 |         # Get the face detector
 67 |         face_detector_module = __import__('face_alignment.detection.' + face_detector,
 68 |                                           globals(), locals(), [face_detector], 0)
 69 |         self.face_detector = face_detector_module.FaceDetector(device=device, verbose=verbose)
 70 | 
 71 |         # Initialise the face alignemnt networks
 72 |         self.face_alignment_net = FAN(network_size)
 73 |         if landmarks_type == LandmarksType._2D:
 74 |             network_name = '2DFAN-' + str(network_size)
 75 |         else:
 76 |             network_name = '3DFAN-' + str(network_size)
 77 | 
 78 |         fan_weights = load_url(models_urls[network_name], map_location=lambda storage, loc: storage)
 79 |         self.face_alignment_net.load_state_dict(fan_weights)
 80 | 
 81 |         self.face_alignment_net.to(device)
 82 |         self.face_alignment_net.eval()
 83 | 
 84 |         # Initialiase the depth prediciton network
 85 |         if landmarks_type == LandmarksType._3D:
 86 |             self.depth_prediciton_net = ResNetDepth()
 87 | 
 88 |             depth_weights = load_url(models_urls['depth'], map_location=lambda storage, loc: storage)
 89 |             depth_dict = {
 90 |                 k.replace('module.', ''): v for k,
 91 |                 v in depth_weights['state_dict'].items()}
 92 |             self.depth_prediciton_net.load_state_dict(depth_dict)
 93 | 
 94 |             self.depth_prediciton_net.to(device)
 95 |             self.depth_prediciton_net.eval()
 96 | 
 97 |     def get_landmarks(self, image_or_path, detected_faces=None):
 98 |         """Deprecated, please use get_landmarks_from_image
 99 | 
100 |         Arguments:
101 |             image_or_path {string or numpy.array or torch.tensor} -- The input image or path to it.
102 | 
103 |         Keyword Arguments:
104 |             detected_faces {list of numpy.array} -- list of bounding boxes, one for each face found
105 |             in the image (default: {None})
106 |         """
107 |         return self.get_landmarks_from_image(image_or_path, detected_faces)
108 | 
109 |     def get_landmarks_from_image(self, image_or_path, detected_faces=None):
110 |         """Predict the landmarks for each face present in the image.
111 | 
112 |         This function predicts a set of 68 2D or 3D images, one for each image present.
113 |         If detect_faces is None the method will also run a face detector.
114 | 
115 |          Arguments:
116 |             image_or_path {string or numpy.array or torch.tensor} -- The input image or path to it.
117 | 
118 |         Keyword Arguments:
119 |             detected_faces {list of numpy.array} -- list of bounding boxes, one for each face found
120 |             in the image (default: {None})
121 |         """
122 |         if isinstance(image_or_path, str):
123 |             try:
124 |                 image = io.imread(image_or_path)
125 |             except IOError:
126 |                 print("error opening file :: ", image_or_path)
127 |                 return None
128 |         else:
129 |             image = image_or_path
130 | 
131 |         if image.ndim == 2:
132 |             image = color.gray2rgb(image)
133 |         elif image.ndim == 4:
134 |             image = image[..., :3]
135 | 
136 |         if detected_faces is None:
137 |             detected_faces = self.face_detector.detect_from_image(image[..., ::-1].copy())
138 | 
139 |         if len(detected_faces) == 0:
140 |             print("Warning: No faces were detected.")
141 |             return None
142 | 
143 |         torch.set_grad_enabled(False)
144 |         landmarks = []
145 |         for i, d in enumerate(detected_faces):
146 |             if (d[2]-d[0])*(d[3]-d[1])<75*75:
147 |                 continue
148 |             center = torch.FloatTensor(
149 |                 [d[2] - (d[2] - d[0]) / 2.0, d[3] - (d[3] - d[1]) / 2.0])
150 |             center[1] = center[1] - (d[3] - d[1]) * 0.12
151 |             scale = (d[2] - d[0] + d[3] - d[1]) / self.face_detector.reference_scale
152 | 
153 |             inp = crop(image, center, scale)
154 |             inp = torch.from_numpy(inp.transpose(
155 |                 (2, 0, 1))).float()
156 | 
157 |             inp = inp.to(self.device)
158 |             inp.div_(255.0).unsqueeze_(0)
159 | 
160 |             out = self.face_alignment_net(inp)[-1].detach()
161 |             if self.flip_input:
162 |                 out += flip(self.face_alignment_net(flip(inp))
163 |                             [-1].detach(), is_label=True)
164 |             out = out.cpu()
165 | 
166 |             pts, pts_img = get_preds_fromhm(out, center, scale)
167 |             pts, pts_img = pts.view(68, 2) * 4, pts_img.view(68, 2)
168 | 
169 |             if self.landmarks_type == LandmarksType._3D:
170 |                 heatmaps = np.zeros((68, 256, 256), dtype=np.float32)
171 |                 for i in range(68):
172 |                     if pts[i, 0] > 0:
173 |                         heatmaps[i] = draw_gaussian(
174 |                             heatmaps[i], pts[i], 2)
175 |                 heatmaps = torch.from_numpy(
176 |                     heatmaps).unsqueeze_(0)
177 | 
178 |                 heatmaps = heatmaps.to(self.device)
179 |                 depth_pred = self.depth_prediciton_net(
180 |                     torch.cat((inp, heatmaps), 1)).data.cpu().view(68, 1)
181 |                 pts_img = torch.cat(
182 |                     (pts_img, depth_pred * (1.0 / (256.0 / (200.0 * scale)))), 1)
183 | 
184 |             landmarks.append(pts_img.numpy())
185 | 
186 |         return landmarks
187 | 
188 |     def get_landmarks_from_directory(self, path, extensions=['.jpg', '.png'], recursive=True, show_progress_bar=True):
189 |         detected_faces = self.face_detector.detect_from_directory(path, extensions, recursive, show_progress_bar)
190 | 
191 |         predictions = {}
192 |         for image_path, bounding_boxes in detected_faces.items():
193 |             image = io.imread(image_path)
194 |             preds = self.get_landmarks_from_image(image, bounding_boxes)
195 |             predictions[image_path] = preds
196 | 
197 |         return predictions
198 | 
199 |     @staticmethod
200 |     def remove_models(self):
201 |         base_path = os.path.join(appdata_dir('face_alignment'), "data")
202 |         for data_model in os.listdir(base_path):
203 |             file_path = os.path.join(base_path, data_model)
204 |             try:
205 |                 if os.path.isfile(file_path):
206 |                     print('Removing ' + data_model + ' ...')
207 |                     os.unlink(file_path)
208 |             except Exception as e:
209 |                 print(e)
210 | 


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/__init__.py:
--------------------------------------------------------------------------------
1 | from .core import FaceDetector


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/__pycache__/__init__.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/face_alignment/detection/__pycache__/__init__.cpython-36.pyc


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/__pycache__/core.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/face_alignment/detection/__pycache__/core.cpython-36.pyc


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/core.py:
--------------------------------------------------------------------------------
  1 | import logging
  2 | import glob
  3 | from tqdm import tqdm
  4 | import numpy as np
  5 | import torch
  6 | import cv2
  7 | from skimage import io
  8 | 
  9 | 
 10 | class FaceDetector(object):
 11 |     """An abstract class representing a face detector.
 12 | 
 13 |     Any other face detection implementation must subclass it. All subclasses
 14 |     must implement ``detect_from_image``, that return a list of detected
 15 |     bounding boxes. Optionally, for speed considerations detect from path is
 16 |     recommended.
 17 |     """
 18 | 
 19 |     def __init__(self, device, verbose):
 20 |         self.device = device
 21 |         self.verbose = verbose
 22 | 
 23 |         if verbose:
 24 |             if 'cpu' in device:
 25 |                 logger = logging.getLogger(__name__)
 26 |                 logger.warning("Detection running on CPU, this may be potentially slow.")
 27 | 
 28 |         if 'cpu' not in device and 'cuda' not in device:
 29 |             if verbose:
 30 |                 logger.error("Expected values for device are: {cpu, cuda} but got: %s", device)
 31 |             raise ValueError
 32 | 
 33 |     def detect_from_image(self, tensor_or_path):
 34 |         """Detects faces in a given image.
 35 | 
 36 |         This function detects the faces present in a provided BGR(usually)
 37 |         image. The input can be either the image itself or the path to it.
 38 | 
 39 |         Arguments:
 40 |             tensor_or_path {numpy.ndarray, torch.tensor or string} -- the path
 41 |             to an image or the image itself.
 42 | 
 43 |         Example::
 44 | 
 45 |             >>> path_to_image = 'data/image_01.jpg'
 46 |             ...   detected_faces = detect_from_image(path_to_image)
 47 |             [A list of bounding boxes (x1, y1, x2, y2)]
 48 |             >>> image = cv2.imread(path_to_image)
 49 |             ...   detected_faces = detect_from_image(image)
 50 |             [A list of bounding boxes (x1, y1, x2, y2)]
 51 | 
 52 |         """
 53 |         raise NotImplementedError
 54 | 
 55 |     def detect_from_directory(self, path, extensions=['.jpg', '.png'], recursive=False, show_progress_bar=True):
 56 |         """Detects faces from all the images present in a given directory.
 57 | 
 58 |         Arguments:
 59 |             path {string} -- a string containing a path that points to the folder containing the images
 60 | 
 61 |         Keyword Arguments:
 62 |             extensions {list} -- list of string containing the extensions to be
 63 |             consider in the following format: ``.extension_name`` (default:
 64 |             {['.jpg', '.png']}) recursive {bool} -- option wherever to scan the
 65 |             folder recursively (default: {False}) show_progress_bar {bool} --
 66 |             display a progressbar (default: {True})
 67 | 
 68 |         Example:
 69 |         >>> directory = 'data'
 70 |         ...   detected_faces = detect_from_directory(directory)
 71 |         {A dictionary of [lists containing bounding boxes(x1, y1, x2, y2)]}
 72 | 
 73 |         """
 74 |         if self.verbose:
 75 |             logger = logging.getLogger(__name__)
 76 | 
 77 |         if len(extensions) == 0:
 78 |             if self.verbose:
 79 |                 logger.error("Expected at list one extension, but none was received.")
 80 |             raise ValueError
 81 | 
 82 |         if self.verbose:
 83 |             logger.info("Constructing the list of images.")
 84 |         additional_pattern = '/**/*' if recursive else '/*'
 85 |         files = []
 86 |         for extension in extensions:
 87 |             files.extend(glob.glob(path + additional_pattern + extension, recursive=recursive))
 88 | 
 89 |         if self.verbose:
 90 |             logger.info("Finished searching for images. %s images found", len(files))
 91 |             logger.info("Preparing to run the detection.")
 92 | 
 93 |         predictions = {}
 94 |         for image_path in tqdm(files, disable=not show_progress_bar):
 95 |             if self.verbose:
 96 |                 logger.info("Running the face detector on image: %s", image_path)
 97 |             predictions[image_path] = self.detect_from_image(image_path)
 98 | 
 99 |         if self.verbose:
100 |             logger.info("The detector was successfully run on all %s images", len(files))
101 | 
102 |         return predictions
103 | 
104 |     @property
105 |     def reference_scale(self):
106 |         raise NotImplementedError
107 | 
108 |     @property
109 |     def reference_x_shift(self):
110 |         raise NotImplementedError
111 | 
112 |     @property
113 |     def reference_y_shift(self):
114 |         raise NotImplementedError
115 | 
116 |     @staticmethod
117 |     def tensor_or_path_to_ndarray(tensor_or_path, rgb=True):
118 |         """Convert path (represented as a string) or torch.tensor to a numpy.ndarray
119 | 
120 |         Arguments:
121 |             tensor_or_path {numpy.ndarray, torch.tensor or string} -- path to the image, or the image itself
122 |         """
123 |         if isinstance(tensor_or_path, str):
124 |             return cv2.imread(tensor_or_path) if not rgb else io.imread(tensor_or_path)
125 |         elif torch.is_tensor(tensor_or_path):
126 |             # Call cpu in case its coming from cuda
127 |             return tensor_or_path.cpu().numpy()[..., ::-1].copy() if not rgb else tensor_or_path.cpu().numpy()
128 |         elif isinstance(tensor_or_path, np.ndarray):
129 |             return tensor_or_path[..., ::-1].copy() if not rgb else tensor_or_path
130 |         else:
131 |             raise TypeError
132 | 


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/dlib/__init__.py:
--------------------------------------------------------------------------------
1 | from .dlib_detector import DlibDetector as FaceDetector


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/dlib/dlib_detector.py:
--------------------------------------------------------------------------------
 1 | import os
 2 | import cv2
 3 | import dlib
 4 | 
 5 | try:
 6 |     import urllib.request as request_file
 7 | except BaseException:
 8 |     import urllib as request_file
 9 | 
10 | from ..core import FaceDetector
11 | from ...utils import appdata_dir
12 | 
13 | 
14 | class DlibDetector(FaceDetector):
15 |     def __init__(self, device, path_to_detector=None, verbose=False):
16 |         super().__init__(device, verbose)
17 | 
18 |         print('Warning: this detector is deprecated. Please use a different one, i.e.: S3FD.')
19 |         base_path = os.path.join(appdata_dir('face_alignment'), "data")
20 | 
21 |         # Initialise the face detector
22 |         if 'cuda' in device:
23 |             if path_to_detector is None:
24 |                 path_to_detector = os.path.join(
25 |                     base_path, "mmod_human_face_detector.dat")
26 | 
27 |                 if not os.path.isfile(path_to_detector):
28 |                     print("Downloading the face detection CNN. Please wait...")
29 | 
30 |                     path_to_temp_detector = os.path.join(
31 |                         base_path, "mmod_human_face_detector.dat.download")
32 | 
33 |                     if os.path.isfile(path_to_temp_detector):
34 |                         os.remove(os.path.join(path_to_temp_detector))
35 | 
36 |                     request_file.urlretrieve(
37 |                         "https://www.adrianbulat.com/downloads/dlib/mmod_human_face_detector.dat",
38 |                         os.path.join(path_to_temp_detector))
39 | 
40 |                     os.rename(os.path.join(path_to_temp_detector), os.path.join(path_to_detector))
41 | 
42 |             self.face_detector = dlib.cnn_face_detection_model_v1(path_to_detector)
43 |         else:
44 |             self.face_detector = dlib.get_frontal_face_detector()
45 | 
46 |     def detect_from_image(self, tensor_or_path):
47 |         image = self.tensor_or_path_to_ndarray(tensor_or_path, rgb=False)
48 | 
49 |         detected_faces = self.face_detector(cv2.cvtColor(image, cv2.COLOR_BGR2GRAY))
50 | 
51 |         if 'cuda' not in self.device:
52 |             detected_faces = [[d.left(), d.top(), d.right(), d.bottom()] for d in detected_faces]
53 |         else:
54 |             detected_faces = [[d.rect.left(), d.rect.top(), d.rect.right(), d.rect.bottom()] for d in detected_faces]
55 | 
56 |         return detected_faces
57 | 
58 |     @property
59 |     def reference_scale(self):
60 |         return 195
61 | 
62 |     @property
63 |     def reference_x_shift(self):
64 |         return 0
65 | 
66 |     @property
67 |     def reference_y_shift(self):
68 |         return 0
69 | 


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/folder/__init__.py:
--------------------------------------------------------------------------------
1 | from .folder_detector import FolderDetector as FaceDetector


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/folder/folder_detector.py:
--------------------------------------------------------------------------------
 1 | import os
 2 | import numpy as np
 3 | import torch
 4 | 
 5 | from ..core import FaceDetector
 6 | 
 7 | 
 8 | class FolderDetector(FaceDetector):
 9 |     '''This is a simple helper module that assumes the faces were detected already
10 |         (either previously or are provided as ground truth).
11 | 
12 |         The class expects to find the bounding boxes in the same format used by
13 |         the rest of face detectors, mainly ``list[(x1,y1,x2,y2),...]``.
14 |         For each image the detector will search for a file with the same name and with one of the
15 |         following extensions: .npy, .t7 or .pth
16 | 
17 |     '''
18 | 
19 |     def __init__(self, device, path_to_detector=None, verbose=False):
20 |         super(FolderDetector, self).__init__(device, verbose)
21 | 
22 |     def detect_from_image(self, tensor_or_path):
23 |         # Only strings supported
24 |         if not isinstance(tensor_or_path, str):
25 |             raise ValueError
26 | 
27 |         base_name = os.path.splitext(tensor_or_path)[0]
28 | 
29 |         if os.path.isfile(base_name + '.npy'):
30 |             detected_faces = np.load(base_name + '.npy')
31 |         elif os.path.isfile(base_name + '.t7'):
32 |             detected_faces = torch.load(base_name + '.t7')
33 |         elif os.path.isfile(base_name + '.pth'):
34 |             detected_faces = torch.load(base_name + '.pth')
35 |         else:
36 |             raise FileNotFoundError
37 | 
38 |         if not isinstance(detected_faces, list):
39 |             raise TypeError
40 | 
41 |         return detected_faces
42 | 
43 |     @property
44 |     def reference_scale(self):
45 |         return 195
46 | 
47 |     @property
48 |     def reference_x_shift(self):
49 |         return 0
50 | 
51 |     @property
52 |     def reference_y_shift(self):
53 |         return 0
54 | 


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/sfd/__init__.py:
--------------------------------------------------------------------------------
1 | from .sfd_detector import SFDDetector as FaceDetector


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/sfd/__pycache__/__init__.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/face_alignment/detection/sfd/__pycache__/__init__.cpython-36.pyc


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/sfd/__pycache__/bbox.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/face_alignment/detection/sfd/__pycache__/bbox.cpython-36.pyc


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/sfd/__pycache__/detect.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/face_alignment/detection/sfd/__pycache__/detect.cpython-36.pyc


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/sfd/__pycache__/net_s3fd.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/face_alignment/detection/sfd/__pycache__/net_s3fd.cpython-36.pyc


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/sfd/__pycache__/sfd_detector.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/face_alignment/detection/sfd/__pycache__/sfd_detector.cpython-36.pyc


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/sfd/bbox.py:
--------------------------------------------------------------------------------
  1 | from __future__ import print_function
  2 | import os
  3 | import sys
  4 | import cv2
  5 | import random
  6 | import datetime
  7 | import time
  8 | import math
  9 | import argparse
 10 | import numpy as np
 11 | import torch
 12 | 
 13 | try:
 14 |     from iou import IOU
 15 | except BaseException:
 16 |     # IOU cython speedup 10x
 17 |     def IOU(ax1, ay1, ax2, ay2, bx1, by1, bx2, by2):
 18 |         sa = abs((ax2 - ax1) * (ay2 - ay1))
 19 |         sb = abs((bx2 - bx1) * (by2 - by1))
 20 |         x1, y1 = max(ax1, bx1), max(ay1, by1)
 21 |         x2, y2 = min(ax2, bx2), min(ay2, by2)
 22 |         w = x2 - x1
 23 |         h = y2 - y1
 24 |         if w < 0 or h < 0:
 25 |             return 0.0
 26 |         else:
 27 |             return 1.0 * w * h / (sa + sb - w * h)
 28 | 
 29 | 
 30 | def bboxlog(x1, y1, x2, y2, axc, ayc, aww, ahh):
 31 |     xc, yc, ww, hh = (x2 + x1) / 2, (y2 + y1) / 2, x2 - x1, y2 - y1
 32 |     dx, dy = (xc - axc) / aww, (yc - ayc) / ahh
 33 |     dw, dh = math.log(ww / aww), math.log(hh / ahh)
 34 |     return dx, dy, dw, dh
 35 | 
 36 | 
 37 | def bboxloginv(dx, dy, dw, dh, axc, ayc, aww, ahh):
 38 |     xc, yc = dx * aww + axc, dy * ahh + ayc
 39 |     ww, hh = math.exp(dw) * aww, math.exp(dh) * ahh
 40 |     x1, x2, y1, y2 = xc - ww / 2, xc + ww / 2, yc - hh / 2, yc + hh / 2
 41 |     return x1, y1, x2, y2
 42 | 
 43 | 
 44 | def nms(dets, thresh):
 45 |     if 0 == len(dets):
 46 |         return []
 47 |     x1, y1, x2, y2, scores = dets[:, 0], dets[:, 1], dets[:, 2], dets[:, 3], dets[:, 4]
 48 |     areas = (x2 - x1 + 1) * (y2 - y1 + 1)
 49 |     order = scores.argsort()[::-1]
 50 | 
 51 |     keep = []
 52 |     while order.size > 0:
 53 |         i = order[0]
 54 |         keep.append(i)
 55 |         xx1, yy1 = np.maximum(x1[i], x1[order[1:]]), np.maximum(y1[i], y1[order[1:]])
 56 |         xx2, yy2 = np.minimum(x2[i], x2[order[1:]]), np.minimum(y2[i], y2[order[1:]])
 57 | 
 58 |         w, h = np.maximum(0.0, xx2 - xx1 + 1), np.maximum(0.0, yy2 - yy1 + 1)
 59 |         ovr = w * h / (areas[i] + areas[order[1:]] - w * h)
 60 | 
 61 |         inds = np.where(ovr <= thresh)[0]
 62 |         order = order[inds + 1]
 63 | 
 64 |     return keep
 65 | 
 66 | 
 67 | def encode(matched, priors, variances):
 68 |     """Encode the variances from the priorbox layers into the ground truth boxes
 69 |     we have matched (based on jaccard overlap) with the prior boxes.
 70 |     Args:
 71 |         matched: (tensor) Coords of ground truth for each prior in point-form
 72 |             Shape: [num_priors, 4].
 73 |         priors: (tensor) Prior boxes in center-offset form
 74 |             Shape: [num_priors,4].
 75 |         variances: (list[float]) Variances of priorboxes
 76 |     Return:
 77 |         encoded boxes (tensor), Shape: [num_priors, 4]
 78 |     """
 79 | 
 80 |     # dist b/t match center and prior's center
 81 |     g_cxcy = (matched[:, :2] + matched[:, 2:]) / 2 - priors[:, :2]
 82 |     # encode variance
 83 |     g_cxcy /= (variances[0] * priors[:, 2:])
 84 |     # match wh / prior wh
 85 |     g_wh = (matched[:, 2:] - matched[:, :2]) / priors[:, 2:]
 86 |     g_wh = torch.log(g_wh) / variances[1]
 87 |     # return target for smooth_l1_loss
 88 |     return torch.cat([g_cxcy, g_wh], 1)  # [num_priors,4]
 89 | 
 90 | 
 91 | def decode(loc, priors, variances):
 92 |     """Decode locations from predictions using priors to undo
 93 |     the encoding we did for offset regression at train time.
 94 |     Args:
 95 |         loc (tensor): location predictions for loc layers,
 96 |             Shape: [num_priors,4]
 97 |         priors (tensor): Prior boxes in center-offset form.
 98 |             Shape: [num_priors,4].
 99 |         variances: (list[float]) Variances of priorboxes
100 |     Return:
101 |         decoded bounding box predictions
102 |     """
103 | 
104 |     boxes = torch.cat((
105 |         priors[:, :2] + loc[:, :2] * variances[0] * priors[:, 2:],
106 |         priors[:, 2:] * torch.exp(loc[:, 2:] * variances[1])), 1)
107 |     boxes[:, :2] -= boxes[:, 2:] / 2
108 |     boxes[:, 2:] += boxes[:, :2]
109 |     return boxes
110 | 


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/sfd/detect.py:
--------------------------------------------------------------------------------
 1 | import torch
 2 | import torch.nn.functional as F
 3 | 
 4 | import os
 5 | import sys
 6 | import cv2
 7 | import random
 8 | import datetime
 9 | import math
10 | import argparse
11 | import numpy as np
12 | 
13 | import scipy.io as sio
14 | import zipfile
15 | from .net_s3fd import s3fd
16 | from .bbox import *
17 | 
18 | 
19 | def detect(net, img, device):
20 |     img = img - np.array([104, 117, 123])
21 |     img = img.transpose(2, 0, 1)
22 |     img = img.reshape((1,) + img.shape)
23 | 
24 |     if 'cuda' in device:
25 |         torch.backends.cudnn.benchmark = True
26 | 
27 |     img = torch.from_numpy(img).float().to(device)
28 |     BB, CC, HH, WW = img.size()
29 |     with torch.no_grad():
30 |         olist = net(img)
31 | 
32 |     bboxlist = []
33 |     for i in range(len(olist) // 2):
34 |         olist[i * 2] = F.softmax(olist[i * 2], dim=1)
35 |     olist = [oelem.data.cpu() for oelem in olist]
36 |     for i in range(len(olist) // 2):
37 |         ocls, oreg = olist[i * 2], olist[i * 2 + 1]
38 |         FB, FC, FH, FW = ocls.size()  # feature map size
39 |         stride = 2**(i + 2)    # 4,8,16,32,64,128
40 |         anchor = stride * 4
41 |         poss = zip(*np.where(ocls[:, 1, :, :] > 0.05))
42 |         for Iindex, hindex, windex in poss:
43 |             axc, ayc = stride / 2 + windex * stride, stride / 2 + hindex * stride
44 |             score = ocls[0, 1, hindex, windex]
45 |             loc = oreg[0, :, hindex, windex].contiguous().view(1, 4)
46 |             priors = torch.Tensor([[axc / 1.0, ayc / 1.0, stride * 4 / 1.0, stride * 4 / 1.0]])
47 |             variances = [0.1, 0.2]
48 |             box = decode(loc, priors, variances)
49 |             x1, y1, x2, y2 = box[0] * 1.0
50 |             # cv2.rectangle(imgshow,(int(x1),int(y1)),(int(x2),int(y2)),(0,0,255),1)
51 |             bboxlist.append([x1, y1, x2, y2, score])
52 |     bboxlist = np.array(bboxlist)
53 |     if 0 == len(bboxlist):
54 |         bboxlist = np.zeros((1, 5))
55 | 
56 |     return bboxlist
57 | 
58 | 
59 | def flip_detect(net, img, device):
60 |     img = cv2.flip(img, 1)
61 |     b = detect(net, img, device)
62 | 
63 |     bboxlist = np.zeros(b.shape)
64 |     bboxlist[:, 0] = img.shape[1] - b[:, 2]
65 |     bboxlist[:, 1] = b[:, 1]
66 |     bboxlist[:, 2] = img.shape[1] - b[:, 0]
67 |     bboxlist[:, 3] = b[:, 3]
68 |     bboxlist[:, 4] = b[:, 4]
69 |     return bboxlist
70 | 
71 | 
72 | def pts_to_bb(pts):
73 |     min_x, min_y = np.min(pts, axis=0)
74 |     max_x, max_y = np.max(pts, axis=0)
75 |     return np.array([min_x, min_y, max_x, max_y])
76 | 


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/sfd/net_s3fd.py:
--------------------------------------------------------------------------------
  1 | import torch
  2 | import torch.nn as nn
  3 | import torch.nn.functional as F
  4 | 
  5 | 
  6 | class L2Norm(nn.Module):
  7 |     def __init__(self, n_channels, scale=1.0):
  8 |         super(L2Norm, self).__init__()
  9 |         self.n_channels = n_channels
 10 |         self.scale = scale
 11 |         self.eps = 1e-10
 12 |         self.weight = nn.Parameter(torch.Tensor(self.n_channels))
 13 |         self.weight.data *= 0.0
 14 |         self.weight.data += self.scale
 15 | 
 16 |     def forward(self, x):
 17 |         norm = x.pow(2).sum(dim=1, keepdim=True).sqrt() + self.eps
 18 |         x = x / norm * self.weight.view(1, -1, 1, 1)
 19 |         return x
 20 | 
 21 | 
 22 | class s3fd(nn.Module):
 23 |     def __init__(self):
 24 |         super(s3fd, self).__init__()
 25 |         self.conv1_1 = nn.Conv2d(3, 64, kernel_size=3, stride=1, padding=1)
 26 |         self.conv1_2 = nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1)
 27 | 
 28 |         self.conv2_1 = nn.Conv2d(64, 128, kernel_size=3, stride=1, padding=1)
 29 |         self.conv2_2 = nn.Conv2d(128, 128, kernel_size=3, stride=1, padding=1)
 30 | 
 31 |         self.conv3_1 = nn.Conv2d(128, 256, kernel_size=3, stride=1, padding=1)
 32 |         self.conv3_2 = nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1)
 33 |         self.conv3_3 = nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1)
 34 | 
 35 |         self.conv4_1 = nn.Conv2d(256, 512, kernel_size=3, stride=1, padding=1)
 36 |         self.conv4_2 = nn.Conv2d(512, 512, kernel_size=3, stride=1, padding=1)
 37 |         self.conv4_3 = nn.Conv2d(512, 512, kernel_size=3, stride=1, padding=1)
 38 | 
 39 |         self.conv5_1 = nn.Conv2d(512, 512, kernel_size=3, stride=1, padding=1)
 40 |         self.conv5_2 = nn.Conv2d(512, 512, kernel_size=3, stride=1, padding=1)
 41 |         self.conv5_3 = nn.Conv2d(512, 512, kernel_size=3, stride=1, padding=1)
 42 | 
 43 |         self.fc6 = nn.Conv2d(512, 1024, kernel_size=3, stride=1, padding=3)
 44 |         self.fc7 = nn.Conv2d(1024, 1024, kernel_size=1, stride=1, padding=0)
 45 | 
 46 |         self.conv6_1 = nn.Conv2d(1024, 256, kernel_size=1, stride=1, padding=0)
 47 |         self.conv6_2 = nn.Conv2d(256, 512, kernel_size=3, stride=2, padding=1)
 48 | 
 49 |         self.conv7_1 = nn.Conv2d(512, 128, kernel_size=1, stride=1, padding=0)
 50 |         self.conv7_2 = nn.Conv2d(128, 256, kernel_size=3, stride=2, padding=1)
 51 | 
 52 |         self.conv3_3_norm = L2Norm(256, scale=10)
 53 |         self.conv4_3_norm = L2Norm(512, scale=8)
 54 |         self.conv5_3_norm = L2Norm(512, scale=5)
 55 | 
 56 |         self.conv3_3_norm_mbox_conf = nn.Conv2d(256, 4, kernel_size=3, stride=1, padding=1)
 57 |         self.conv3_3_norm_mbox_loc = nn.Conv2d(256, 4, kernel_size=3, stride=1, padding=1)
 58 |         self.conv4_3_norm_mbox_conf = nn.Conv2d(512, 2, kernel_size=3, stride=1, padding=1)
 59 |         self.conv4_3_norm_mbox_loc = nn.Conv2d(512, 4, kernel_size=3, stride=1, padding=1)
 60 |         self.conv5_3_norm_mbox_conf = nn.Conv2d(512, 2, kernel_size=3, stride=1, padding=1)
 61 |         self.conv5_3_norm_mbox_loc = nn.Conv2d(512, 4, kernel_size=3, stride=1, padding=1)
 62 | 
 63 |         self.fc7_mbox_conf = nn.Conv2d(1024, 2, kernel_size=3, stride=1, padding=1)
 64 |         self.fc7_mbox_loc = nn.Conv2d(1024, 4, kernel_size=3, stride=1, padding=1)
 65 |         self.conv6_2_mbox_conf = nn.Conv2d(512, 2, kernel_size=3, stride=1, padding=1)
 66 |         self.conv6_2_mbox_loc = nn.Conv2d(512, 4, kernel_size=3, stride=1, padding=1)
 67 |         self.conv7_2_mbox_conf = nn.Conv2d(256, 2, kernel_size=3, stride=1, padding=1)
 68 |         self.conv7_2_mbox_loc = nn.Conv2d(256, 4, kernel_size=3, stride=1, padding=1)
 69 | 
 70 |     def forward(self, x):
 71 |         h = F.relu(self.conv1_1(x))
 72 |         h = F.relu(self.conv1_2(h))
 73 |         h = F.max_pool2d(h, 2, 2)
 74 | 
 75 |         h = F.relu(self.conv2_1(h))
 76 |         h = F.relu(self.conv2_2(h))
 77 |         h = F.max_pool2d(h, 2, 2)
 78 | 
 79 |         h = F.relu(self.conv3_1(h))
 80 |         h = F.relu(self.conv3_2(h))
 81 |         h = F.relu(self.conv3_3(h))
 82 |         f3_3 = h
 83 |         h = F.max_pool2d(h, 2, 2)
 84 | 
 85 |         h = F.relu(self.conv4_1(h))
 86 |         h = F.relu(self.conv4_2(h))
 87 |         h = F.relu(self.conv4_3(h))
 88 |         f4_3 = h
 89 |         h = F.max_pool2d(h, 2, 2)
 90 | 
 91 |         h = F.relu(self.conv5_1(h))
 92 |         h = F.relu(self.conv5_2(h))
 93 |         h = F.relu(self.conv5_3(h))
 94 |         f5_3 = h
 95 |         h = F.max_pool2d(h, 2, 2)
 96 | 
 97 |         h = F.relu(self.fc6(h))
 98 |         h = F.relu(self.fc7(h))
 99 |         ffc7 = h
100 |         h = F.relu(self.conv6_1(h))
101 |         h = F.relu(self.conv6_2(h))
102 |         f6_2 = h
103 |         h = F.relu(self.conv7_1(h))
104 |         h = F.relu(self.conv7_2(h))
105 |         f7_2 = h
106 | 
107 |         f3_3 = self.conv3_3_norm(f3_3)
108 |         f4_3 = self.conv4_3_norm(f4_3)
109 |         f5_3 = self.conv5_3_norm(f5_3)
110 | 
111 |         cls1 = self.conv3_3_norm_mbox_conf(f3_3)
112 |         reg1 = self.conv3_3_norm_mbox_loc(f3_3)
113 |         cls2 = self.conv4_3_norm_mbox_conf(f4_3)
114 |         reg2 = self.conv4_3_norm_mbox_loc(f4_3)
115 |         cls3 = self.conv5_3_norm_mbox_conf(f5_3)
116 |         reg3 = self.conv5_3_norm_mbox_loc(f5_3)
117 |         cls4 = self.fc7_mbox_conf(ffc7)
118 |         reg4 = self.fc7_mbox_loc(ffc7)
119 |         cls5 = self.conv6_2_mbox_conf(f6_2)
120 |         reg5 = self.conv6_2_mbox_loc(f6_2)
121 |         cls6 = self.conv7_2_mbox_conf(f7_2)
122 |         reg6 = self.conv7_2_mbox_loc(f7_2)
123 | 
124 |         # max-out background label
125 |         chunk = torch.chunk(cls1, 4, 1)
126 |         bmax = torch.max(torch.max(chunk[0], chunk[1]), chunk[2])
127 |         cls1 = torch.cat([bmax, chunk[3]], dim=1)
128 | 
129 |         return [cls1, reg1, cls2, reg2, cls3, reg3, cls4, reg4, cls5, reg5, cls6, reg6]
130 | 


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/detection/sfd/sfd_detector.py:
--------------------------------------------------------------------------------
 1 | import os
 2 | import cv2
 3 | from torch.utils.model_zoo import load_url
 4 | 
 5 | from ..core import FaceDetector
 6 | 
 7 | from .net_s3fd import s3fd
 8 | from .bbox import *
 9 | from .detect import *
10 | 
11 | models_urls = {
12 |     's3fd': 'https://www.adrianbulat.com/downloads/python-fan/s3fd-619a316812.pth',
13 | }
14 | 
15 | 
16 | class SFDDetector(FaceDetector):
17 |     def __init__(self, device, path_to_detector=None, verbose=False):
18 |         super(SFDDetector, self).__init__(device, verbose)
19 | 
20 |         # Initialise the face detector
21 |         if path_to_detector is None:
22 |             model_weights = load_url(models_urls['s3fd'])
23 |         else:
24 |             model_weights = torch.load(path_to_detector)
25 | 
26 |         self.face_detector = s3fd()
27 |         self.face_detector.load_state_dict(model_weights)
28 |         self.face_detector.to(device)
29 |         self.face_detector.eval()
30 | 
31 |     def detect_from_image(self, tensor_or_path):
32 |         image = self.tensor_or_path_to_ndarray(tensor_or_path)
33 | 
34 |         bboxlist = detect(self.face_detector, image, device=self.device)
35 |         keep = nms(bboxlist, 0.3)
36 |         bboxlist = bboxlist[keep, :]
37 |         bboxlist = [x for x in bboxlist if x[-1] > 0.5]
38 | 
39 |         return bboxlist
40 | 
41 |     @property
42 |     def reference_scale(self):
43 |         return 195
44 | 
45 |     @property
46 |     def reference_x_shift(self):
47 |         return 0
48 | 
49 |     @property
50 |     def reference_y_shift(self):
51 |         return 0
52 | 


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/models.py:
--------------------------------------------------------------------------------
  1 | import torch
  2 | import torch.nn as nn
  3 | import torch.nn.functional as F
  4 | import math
  5 | 
  6 | 
  7 | def conv3x3(in_planes, out_planes, strd=1, padding=1, bias=False):
  8 |     "3x3 convolution with padding"
  9 |     return nn.Conv2d(in_planes, out_planes, kernel_size=3,
 10 |                      stride=strd, padding=padding, bias=bias)
 11 | 
 12 | 
 13 | class ConvBlock(nn.Module):
 14 |     def __init__(self, in_planes, out_planes):
 15 |         super(ConvBlock, self).__init__()
 16 |         self.bn1 = nn.BatchNorm2d(in_planes)
 17 |         self.conv1 = conv3x3(in_planes, int(out_planes / 2))
 18 |         self.bn2 = nn.BatchNorm2d(int(out_planes / 2))
 19 |         self.conv2 = conv3x3(int(out_planes / 2), int(out_planes / 4))
 20 |         self.bn3 = nn.BatchNorm2d(int(out_planes / 4))
 21 |         self.conv3 = conv3x3(int(out_planes / 4), int(out_planes / 4))
 22 | 
 23 |         if in_planes != out_planes:
 24 |             self.downsample = nn.Sequential(
 25 |                 nn.BatchNorm2d(in_planes),
 26 |                 nn.ReLU(True),
 27 |                 nn.Conv2d(in_planes, out_planes,
 28 |                           kernel_size=1, stride=1, bias=False),
 29 |             )
 30 |         else:
 31 |             self.downsample = None
 32 | 
 33 |     def forward(self, x):
 34 |         residual = x
 35 | 
 36 |         out1 = self.bn1(x)
 37 |         out1 = F.relu(out1, True)
 38 |         out1 = self.conv1(out1)
 39 | 
 40 |         out2 = self.bn2(out1)
 41 |         out2 = F.relu(out2, True)
 42 |         out2 = self.conv2(out2)
 43 | 
 44 |         out3 = self.bn3(out2)
 45 |         out3 = F.relu(out3, True)
 46 |         out3 = self.conv3(out3)
 47 | 
 48 |         out3 = torch.cat((out1, out2, out3), 1)
 49 | 
 50 |         if self.downsample is not None:
 51 |             residual = self.downsample(residual)
 52 | 
 53 |         out3 += residual
 54 | 
 55 |         return out3
 56 | 
 57 | 
 58 | class Bottleneck(nn.Module):
 59 | 
 60 |     expansion = 4
 61 | 
 62 |     def __init__(self, inplanes, planes, stride=1, downsample=None):
 63 |         super(Bottleneck, self).__init__()
 64 |         self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1, bias=False)
 65 |         self.bn1 = nn.BatchNorm2d(planes)
 66 |         self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride,
 67 |                                padding=1, bias=False)
 68 |         self.bn2 = nn.BatchNorm2d(planes)
 69 |         self.conv3 = nn.Conv2d(planes, planes * 4, kernel_size=1, bias=False)
 70 |         self.bn3 = nn.BatchNorm2d(planes * 4)
 71 |         self.relu = nn.ReLU(inplace=True)
 72 |         self.downsample = downsample
 73 |         self.stride = stride
 74 | 
 75 |     def forward(self, x):
 76 |         residual = x
 77 | 
 78 |         out = self.conv1(x)
 79 |         out = self.bn1(out)
 80 |         out = self.relu(out)
 81 | 
 82 |         out = self.conv2(out)
 83 |         out = self.bn2(out)
 84 |         out = self.relu(out)
 85 | 
 86 |         out = self.conv3(out)
 87 |         out = self.bn3(out)
 88 | 
 89 |         if self.downsample is not None:
 90 |             residual = self.downsample(x)
 91 | 
 92 |         out += residual
 93 |         out = self.relu(out)
 94 | 
 95 |         return out
 96 | 
 97 | 
 98 | class HourGlass(nn.Module):
 99 |     def __init__(self, num_modules, depth, num_features):
100 |         super(HourGlass, self).__init__()
101 |         self.num_modules = num_modules
102 |         self.depth = depth
103 |         self.features = num_features
104 | 
105 |         self._generate_network(self.depth)
106 | 
107 |     def _generate_network(self, level):
108 |         self.add_module('b1_' + str(level), ConvBlock(self.features, self.features))
109 | 
110 |         self.add_module('b2_' + str(level), ConvBlock(self.features, self.features))
111 | 
112 |         if level > 1:
113 |             self._generate_network(level - 1)
114 |         else:
115 |             self.add_module('b2_plus_' + str(level), ConvBlock(self.features, self.features))
116 | 
117 |         self.add_module('b3_' + str(level), ConvBlock(self.features, self.features))
118 | 
119 |     def _forward(self, level, inp):
120 |         # Upper branch
121 |         up1 = inp
122 |         up1 = self._modules['b1_' + str(level)](up1)
123 | 
124 |         # Lower branch
125 |         low1 = F.avg_pool2d(inp, 2, stride=2)
126 |         low1 = self._modules['b2_' + str(level)](low1)
127 | 
128 |         if level > 1:
129 |             low2 = self._forward(level - 1, low1)
130 |         else:
131 |             low2 = low1
132 |             low2 = self._modules['b2_plus_' + str(level)](low2)
133 | 
134 |         low3 = low2
135 |         low3 = self._modules['b3_' + str(level)](low3)
136 | 
137 |         up2 = F.interpolate(low3, scale_factor=2, mode='nearest')
138 | 
139 |         return up1 + up2
140 | 
141 |     def forward(self, x):
142 |         return self._forward(self.depth, x)
143 | 
144 | 
145 | class FAN(nn.Module):
146 | 
147 |     def __init__(self, num_modules=1):
148 |         super(FAN, self).__init__()
149 |         self.num_modules = num_modules
150 | 
151 |         # Base part
152 |         self.conv1 = nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3)
153 |         self.bn1 = nn.BatchNorm2d(64)
154 |         self.conv2 = ConvBlock(64, 128)
155 |         self.conv3 = ConvBlock(128, 128)
156 |         self.conv4 = ConvBlock(128, 256)
157 | 
158 |         # Stacking part
159 |         for hg_module in range(self.num_modules):
160 |             self.add_module('m' + str(hg_module), HourGlass(1, 4, 256))
161 |             self.add_module('top_m_' + str(hg_module), ConvBlock(256, 256))
162 |             self.add_module('conv_last' + str(hg_module),
163 |                             nn.Conv2d(256, 256, kernel_size=1, stride=1, padding=0))
164 |             self.add_module('bn_end' + str(hg_module), nn.BatchNorm2d(256))
165 |             self.add_module('l' + str(hg_module), nn.Conv2d(256,
166 |                                                             68, kernel_size=1, stride=1, padding=0))
167 | 
168 |             if hg_module < self.num_modules - 1:
169 |                 self.add_module(
170 |                     'bl' + str(hg_module), nn.Conv2d(256, 256, kernel_size=1, stride=1, padding=0))
171 |                 self.add_module('al' + str(hg_module), nn.Conv2d(68,
172 |                                                                  256, kernel_size=1, stride=1, padding=0))
173 | 
174 |     def forward(self, x):
175 |         x = F.relu(self.bn1(self.conv1(x)), True)
176 |         x = F.avg_pool2d(self.conv2(x), 2, stride=2)
177 |         x = self.conv3(x)
178 |         x = self.conv4(x)
179 | 
180 |         previous = x
181 | 
182 |         outputs = []
183 |         for i in range(self.num_modules):
184 |             hg = self._modules['m' + str(i)](previous)
185 | 
186 |             ll = hg
187 |             ll = self._modules['top_m_' + str(i)](ll)
188 | 
189 |             ll = F.relu(self._modules['bn_end' + str(i)]
190 |                         (self._modules['conv_last' + str(i)](ll)), True)
191 | 
192 |             # Predict heatmaps
193 |             tmp_out = self._modules['l' + str(i)](ll)
194 |             outputs.append(tmp_out)
195 | 
196 |             if i < self.num_modules - 1:
197 |                 ll = self._modules['bl' + str(i)](ll)
198 |                 tmp_out_ = self._modules['al' + str(i)](tmp_out)
199 |                 previous = previous + ll + tmp_out_
200 | 
201 |         return outputs
202 | 
203 | 
204 | class ResNetDepth(nn.Module):
205 | 
206 |     def __init__(self, block=Bottleneck, layers=[3, 8, 36, 3], num_classes=68):
207 |         self.inplanes = 64
208 |         super(ResNetDepth, self).__init__()
209 |         self.conv1 = nn.Conv2d(3 + 68, 64, kernel_size=7, stride=2, padding=3,
210 |                                bias=False)
211 |         self.bn1 = nn.BatchNorm2d(64)
212 |         self.relu = nn.ReLU(inplace=True)
213 |         self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
214 |         self.layer1 = self._make_layer(block, 64, layers[0])
215 |         self.layer2 = self._make_layer(block, 128, layers[1], stride=2)
216 |         self.layer3 = self._make_layer(block, 256, layers[2], stride=2)
217 |         self.layer4 = self._make_layer(block, 512, layers[3], stride=2)
218 |         self.avgpool = nn.AvgPool2d(7)
219 |         self.fc = nn.Linear(512 * block.expansion, num_classes)
220 | 
221 |         for m in self.modules():
222 |             if isinstance(m, nn.Conv2d):
223 |                 n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
224 |                 m.weight.data.normal_(0, math.sqrt(2. / n))
225 |             elif isinstance(m, nn.BatchNorm2d):
226 |                 m.weight.data.fill_(1)
227 |                 m.bias.data.zero_()
228 | 
229 |     def _make_layer(self, block, planes, blocks, stride=1):
230 |         downsample = None
231 |         if stride != 1 or self.inplanes != planes * block.expansion:
232 |             downsample = nn.Sequential(
233 |                 nn.Conv2d(self.inplanes, planes * block.expansion,
234 |                           kernel_size=1, stride=stride, bias=False),
235 |                 nn.BatchNorm2d(planes * block.expansion),
236 |             )
237 | 
238 |         layers = []
239 |         layers.append(block(self.inplanes, planes, stride, downsample))
240 |         self.inplanes = planes * block.expansion
241 |         for i in range(1, blocks):
242 |             layers.append(block(self.inplanes, planes))
243 | 
244 |         return nn.Sequential(*layers)
245 | 
246 |     def forward(self, x):
247 |         x = self.conv1(x)
248 |         x = self.bn1(x)
249 |         x = self.relu(x)
250 |         x = self.maxpool(x)
251 | 
252 |         x = self.layer1(x)
253 |         x = self.layer2(x)
254 |         x = self.layer3(x)
255 |         x = self.layer4(x)
256 | 
257 |         x = self.avgpool(x)
258 |         x = x.view(x.size(0), -1)
259 |         x = self.fc(x)
260 | 
261 |         return x
262 | 


--------------------------------------------------------------------------------
/face-alignment_projection/face_alignment/utils.py:
--------------------------------------------------------------------------------
  1 | from __future__ import print_function
  2 | import os
  3 | import sys
  4 | import time
  5 | import torch
  6 | import math
  7 | import numpy as np
  8 | import cv2
  9 | 
 10 | 
 11 | def _gaussian(
 12 |         size=3, sigma=0.25, amplitude=1, normalize=False, width=None,
 13 |         height=None, sigma_horz=None, sigma_vert=None, mean_horz=0.5,
 14 |         mean_vert=0.5):
 15 |     # handle some defaults
 16 |     if width is None:
 17 |         width = size
 18 |     if height is None:
 19 |         height = size
 20 |     if sigma_horz is None:
 21 |         sigma_horz = sigma
 22 |     if sigma_vert is None:
 23 |         sigma_vert = sigma
 24 |     center_x = mean_horz * width + 0.5
 25 |     center_y = mean_vert * height + 0.5
 26 |     gauss = np.empty((height, width), dtype=np.float32)
 27 |     # generate kernel
 28 |     for i in range(height):
 29 |         for j in range(width):
 30 |             gauss[i][j] = amplitude * math.exp(-(math.pow((j + 1 - center_x) / (
 31 |                 sigma_horz * width), 2) / 2.0 + math.pow((i + 1 - center_y) / (sigma_vert * height), 2) / 2.0))
 32 |     if normalize:
 33 |         gauss = gauss / np.sum(gauss)
 34 |     return gauss
 35 | 
 36 | 
 37 | def draw_gaussian(image, point, sigma):
 38 |     # Check if the gaussian is inside
 39 |     ul = [math.floor(point[0] - 3 * sigma), math.floor(point[1] - 3 * sigma)]
 40 |     br = [math.floor(point[0] + 3 * sigma), math.floor(point[1] + 3 * sigma)]
 41 |     if (ul[0] > image.shape[1] or ul[1] > image.shape[0] or br[0] < 1 or br[1] < 1):
 42 |         return image
 43 |     size = 6 * sigma + 1
 44 |     g = _gaussian(size)
 45 |     g_x = [int(max(1, -ul[0])), int(min(br[0], image.shape[1])) - int(max(1, ul[0])) + int(max(1, -ul[0]))]
 46 |     g_y = [int(max(1, -ul[1])), int(min(br[1], image.shape[0])) - int(max(1, ul[1])) + int(max(1, -ul[1]))]
 47 |     img_x = [int(max(1, ul[0])), int(min(br[0], image.shape[1]))]
 48 |     img_y = [int(max(1, ul[1])), int(min(br[1], image.shape[0]))]
 49 |     assert (g_x[0] > 0 and g_y[1] > 0)
 50 |     image[img_y[0] - 1:img_y[1], img_x[0] - 1:img_x[1]
 51 |           ] = image[img_y[0] - 1:img_y[1], img_x[0] - 1:img_x[1]] + g[g_y[0] - 1:g_y[1], g_x[0] - 1:g_x[1]]
 52 |     image[image > 1] = 1
 53 |     return image
 54 | 
 55 | 
 56 | def transform(point, center, scale, resolution, invert=False):
 57 |     """Generate and affine transformation matrix.
 58 | 
 59 |     Given a set of points, a center, a scale and a targer resolution, the
 60 |     function generates and affine transformation matrix. If invert is ``True``
 61 |     it will produce the inverse transformation.
 62 | 
 63 |     Arguments:
 64 |         point {torch.tensor} -- the input 2D point
 65 |         center {torch.tensor or numpy.array} -- the center around which to perform the transformations
 66 |         scale {float} -- the scale of the face/object
 67 |         resolution {float} -- the output resolution
 68 | 
 69 |     Keyword Arguments:
 70 |         invert {bool} -- define wherever the function should produce the direct or the
 71 |         inverse transformation matrix (default: {False})
 72 |     """
 73 |     _pt = torch.ones(3)
 74 |     _pt[0] = point[0]
 75 |     _pt[1] = point[1]
 76 | 
 77 |     h = 200.0 * scale
 78 |     t = torch.eye(3)
 79 |     t[0, 0] = resolution / h
 80 |     t[1, 1] = resolution / h
 81 |     t[0, 2] = resolution * (-center[0] / h + 0.5)
 82 |     t[1, 2] = resolution * (-center[1] / h + 0.5)
 83 | 
 84 |     if invert:
 85 |         t = torch.inverse(t)
 86 | 
 87 |     new_point = (torch.matmul(t, _pt))[0:2]
 88 | 
 89 |     return new_point.int()
 90 | 
 91 | 
 92 | def crop(image, center, scale, resolution=256.0):
 93 |     """Center crops an image or set of heatmaps
 94 | 
 95 |     Arguments:
 96 |         image {numpy.array} -- an rgb image
 97 |         center {numpy.array} -- the center of the object, usually the same as of the bounding box
 98 |         scale {float} -- scale of the face
 99 | 
100 |     Keyword Arguments:
101 |         resolution {float} -- the size of the output cropped image (default: {256.0})
102 | 
103 |     Returns:
104 |         [type] -- [description]
105 |     """  # Crop around the center point
106 |     """ Crops the image around the center. Input is expected to be an np.ndarray """
107 |     ul = transform([1, 1], center, scale, resolution, True)
108 |     br = transform([resolution, resolution], center, scale, resolution, True)
109 |     # pad = math.ceil(torch.norm((ul - br).float()) / 2.0 - (br[0] - ul[0]) / 2.0)
110 |     if image.ndim > 2:
111 |         newDim = np.array([br[1] - ul[1], br[0] - ul[0],
112 |                            image.shape[2]], dtype=np.int32)
113 |         newImg = np.zeros(newDim, dtype=np.uint8)
114 |     else:
115 |         newDim = np.array([br[1] - ul[1], br[0] - ul[0]], dtype=np.int)
116 |         newImg = np.zeros(newDim, dtype=np.uint8)
117 |     ht = image.shape[0]
118 |     wd = image.shape[1]
119 |     newX = np.array(
120 |         [max(1, -ul[0] + 1), min(br[0], wd) - ul[0]], dtype=np.int32)
121 |     newY = np.array(
122 |         [max(1, -ul[1] + 1), min(br[1], ht) - ul[1]], dtype=np.int32)
123 |     oldX = np.array([max(1, ul[0] + 1), min(br[0], wd)], dtype=np.int32)
124 |     oldY = np.array([max(1, ul[1] + 1), min(br[1], ht)], dtype=np.int32)
125 |     newImg[newY[0] - 1:newY[1], newX[0] - 1:newX[1]
126 |            ] = image[oldY[0] - 1:oldY[1], oldX[0] - 1:oldX[1], :]
127 |     newImg = cv2.resize(newImg, dsize=(int(resolution), int(resolution)),
128 |                         interpolation=cv2.INTER_LINEAR)
129 |     return newImg
130 | 
131 | 
132 | def get_preds_fromhm(hm, center=None, scale=None):
133 |     """Obtain (x,y) coordinates given a set of N heatmaps. If the center
134 |     and the scale is provided the function will return the points also in
135 |     the original coordinate frame.
136 | 
137 |     Arguments:
138 |         hm {torch.tensor} -- the predicted heatmaps, of shape [B, N, W, H]
139 | 
140 |     Keyword Arguments:
141 |         center {torch.tensor} -- the center of the bounding box (default: {None})
142 |         scale {float} -- face scale (default: {None})
143 |     """
144 |     max, idx = torch.max(
145 |         hm.view(hm.size(0), hm.size(1), hm.size(2) * hm.size(3)), 2)
146 |     idx += 1
147 |     preds = idx.view(idx.size(0), idx.size(1), 1).repeat(1, 1, 2).float()
148 |     preds[..., 0].apply_(lambda x: (x - 1) % hm.size(3) + 1)
149 |     preds[..., 1].add_(-1).div_(hm.size(2)).floor_().add_(1)
150 | 
151 |     for i in range(preds.size(0)):
152 |         for j in range(preds.size(1)):
153 |             hm_ = hm[i, j, :]
154 |             pX, pY = int(preds[i, j, 0]) - 1, int(preds[i, j, 1]) - 1
155 |             if pX > 0 and pX < 63 and pY > 0 and pY < 63:
156 |                 diff = torch.FloatTensor(
157 |                     [hm_[pY, pX + 1] - hm_[pY, pX - 1],
158 |                      hm_[pY + 1, pX] - hm_[pY - 1, pX]])
159 |                 preds[i, j].add_(diff.sign_().mul_(.25))
160 | 
161 |     preds.add_(-.5)
162 | 
163 |     preds_orig = torch.zeros(preds.size())
164 |     if center is not None and scale is not None:
165 |         for i in range(hm.size(0)):
166 |             for j in range(hm.size(1)):
167 |                 preds_orig[i, j] = transform(
168 |                     preds[i, j], center, scale, hm.size(2), True)
169 | 
170 |     return preds, preds_orig
171 | 
172 | 
173 | def shuffle_lr(parts, pairs=None):
174 |     """Shuffle the points left-right according to the axis of symmetry
175 |     of the object.
176 | 
177 |     Arguments:
178 |         parts {torch.tensor} -- a 3D or 4D object containing the
179 |         heatmaps.
180 | 
181 |     Keyword Arguments:
182 |         pairs {list of integers} -- [order of the flipped points] (default: {None})
183 |     """
184 |     if pairs is None:
185 |         pairs = [16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0,
186 |                  26, 25, 24, 23, 22, 21, 20, 19, 18, 17, 27, 28, 29, 30, 35,
187 |                  34, 33, 32, 31, 45, 44, 43, 42, 47, 46, 39, 38, 37, 36, 41,
188 |                  40, 54, 53, 52, 51, 50, 49, 48, 59, 58, 57, 56, 55, 64, 63,
189 |                  62, 61, 60, 67, 66, 65]
190 |     if parts.ndimension() == 3:
191 |         parts = parts[pairs, ...]
192 |     else:
193 |         parts = parts[:, pairs, ...]
194 | 
195 |     return parts
196 | 
197 | 
198 | def flip(tensor, is_label=False):
199 |     """Flip an image or a set of heatmaps left-right
200 | 
201 |     Arguments:
202 |         tensor {numpy.array or torch.tensor} -- [the input image or heatmaps]
203 | 
204 |     Keyword Arguments:
205 |         is_label {bool} -- [denote wherever the input is an image or a set of heatmaps ] (default: {False})
206 |     """
207 |     if not torch.is_tensor(tensor):
208 |         tensor = torch.from_numpy(tensor)
209 | 
210 |     if is_label:
211 |         tensor = shuffle_lr(tensor).flip(tensor.ndimension() - 1)
212 |     else:
213 |         tensor = tensor.flip(tensor.ndimension() - 1)
214 | 
215 |     return tensor
216 | 
217 | # From pyzolib/paths.py (https://bitbucket.org/pyzo/pyzolib/src/tip/paths.py)
218 | 
219 | 
220 | def appdata_dir(appname=None, roaming=False):
221 |     """ appdata_dir(appname=None, roaming=False)
222 | 
223 |     Get the path to the application directory, where applications are allowed
224 |     to write user specific files (e.g. configurations). For non-user specific
225 |     data, consider using common_appdata_dir().
226 |     If appname is given, a subdir is appended (and created if necessary).
227 |     If roaming is True, will prefer a roaming directory (Windows Vista/7).
228 |     """
229 | 
230 |     # Define default user directory
231 |     userDir = os.getenv('FACEALIGNMENT_USERDIR', None)
232 |     if userDir is None:
233 |         userDir = os.path.expanduser('~')
234 |         if not os.path.isdir(userDir):  # pragma: no cover
235 |             userDir = '/var/tmp'  # issue #54
236 | 
237 |     # Get system app data dir
238 |     path = None
239 |     if sys.platform.startswith('win'):
240 |         path1, path2 = os.getenv('LOCALAPPDATA'), os.getenv('APPDATA')
241 |         path = (path2 or path1) if roaming else (path1 or path2)
242 |     elif sys.platform.startswith('darwin'):
243 |         path = os.path.join(userDir, 'Library', 'Application Support')
244 |     # On Linux and as fallback
245 |     if not (path and os.path.isdir(path)):
246 |         path = userDir
247 | 
248 |     # Maybe we should store things local to the executable (in case of a
249 |     # portable distro or a frozen application that wants to be portable)
250 |     prefix = sys.prefix
251 |     if getattr(sys, 'frozen', None):
252 |         prefix = os.path.abspath(os.path.dirname(sys.executable))
253 |     for reldir in ('settings', '../settings'):
254 |         localpath = os.path.abspath(os.path.join(prefix, reldir))
255 |         if os.path.isdir(localpath):  # pragma: no cover
256 |             try:
257 |                 open(os.path.join(localpath, 'test.write'), 'wb').close()
258 |                 os.remove(os.path.join(localpath, 'test.write'))
259 |             except IOError:
260 |                 pass  # We cannot write in this directory
261 |             else:
262 |                 path = localpath
263 |                 break
264 | 
265 |     # Get path specific for this app
266 |     if appname:
267 |         if path == userDir:
268 |             appname = '.' + appname.lstrip('.')  # Make it a hidden directory
269 |         path = os.path.join(path, appname)
270 |         if not os.path.isdir(path):  # pragma: no cover
271 |             os.mkdir(path)
272 | 
273 |     # Done
274 |     return path
275 | 


--------------------------------------------------------------------------------
/face-alignment_projection/get_landmarks.py:
--------------------------------------------------------------------------------
 1 | import cv2
 2 | import numpy
 3 | import face_alignment
 4 | import numpy as np
 5 | import skimage.io as io
 6 | import os
 7 | 
 8 | fa = face_alignment.FaceAlignment(face_alignment.LandmarksType._2D, flip_input=True)
 9 | predefined_color = [128,64,160,244,35,220,80,200]
10 | 
11 | dataset = '../../dataset/voxcelb1/data'
12 | person_name = os.listdir(dataset)
13 | num_classes = len(person_name)
14 | video_path = [os.path.abspath(os.path.join(dataset,i)) for i in person_name]
15 | 
16 | # save all video clips
17 | all_video_path = []
18 | f = open('video_clips_path.txt','w')
19 | print('saving video path!!')
20 | for v in video_path:
21 |     each_person_clip = os.path.join(v,'1.6')
22 |     each_person_clip = [ os.path.join(each_person_clip,i,'1') for i in os.listdir(each_person_clip)]
23 |     all_video_path += each_person_clip
24 |     for path in each_person_clip:
25 |         print(path)
26 |         f.write(path+'\n')
27 | f.close()
28 | if not os.path.exists('../../dataset/voxcelb1/landmarks'):
29 |     os.makedirs('../../dataset/voxcelb1/landmarks')
30 | 
31 | print('saving landmarks!!!')
32 | f = open('video_landmarks_path.txt','w')
33 | for v in all_video_path:
34 |     print(v)
35 |     for i in os.listdir(v):
36 |         img = os.path.join(v,i)
37 |         basename = os.path.dirname(img)
38 |         filename = os.path.basename(img)
39 |         input = io.imread(img)
40 |         preds = fa.get_landmarks(input)
41 |         if preds is None:
42 |             continue
43 |         preds = preds[-1]
44 |         saved = np.ones_like(input,dtype=np.uint8)*255
45 |         for i in range(17-1):
46 |             cv2.line(saved,(preds[i,0],preds[i,1]),(preds[i+1,0],preds[i+1,1]),[predefined_color[0],250,10],2)
47 |         for i in range(17,22-1):
48 |             cv2.line(saved,(preds[i,0],preds[i,1]),(preds[i+1,0],preds[i+1,1]),[predefined_color[1],10,250],2)
49 |         for i in range(22,27-1):
50 |             cv2.line(saved,(preds[i,0],preds[i,1]),(preds[i+1,0],preds[i+1,1]),[predefined_color[2],150,150],2)
51 |         for i in range(27,31-1):
52 |             cv2.line(saved, (preds[i, 0], preds[i, 1]), (preds[i + 1, 0], preds[i + 1, 1]), [predefined_color[3],0,0],
53 |                      2)
54 |         for i in range(31,36-1):
55 |             cv2.line(saved,(preds[i,0],preds[i,1]),(preds[i+1,0],preds[i+1,1]),[predefined_color[4],50,0],2)
56 |         for i in range(36,42-1):
57 |             cv2.line(saved,(preds[i,0],preds[i,1]),(preds[i+1,0],preds[i+1,1]),[predefined_color[5],0,50],2)
58 |         for i in range(42,48-1):
59 |             cv2.line(saved,(preds[i,0],preds[i,1]),(preds[i+1,0],preds[i+1,1]),[predefined_color[6],180,30],2)
60 |         for i in range(48,60-1):
61 |             cv2.line(saved, (preds[i, 0], preds[i, 1]), (preds[i + 1, 0], preds[i + 1, 1]), [predefined_color[7],20,180],
62 |                      2)
63 |         basename = basename.split('data')
64 |         basename = basename[0] + 'dataset/voxcelb1/' 'landmarks' + basename[2]
65 |         if not os.path.exists(basename):
66 |             os.makedirs(basename)
67 |         filename = os.path.join(basename,filename)
68 |         print(filename)
69 |         f.write(filename + '\n')
70 |         cv2.imwrite(filename,saved)
71 | f.close()
72 | 
73 | 
74 | 


--------------------------------------------------------------------------------
/face-alignment_projection/songyiren.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/songyiren.jpg


--------------------------------------------------------------------------------
/face-alignment_projection/test/assets/aflw-test.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/face-alignment_projection/test/assets/aflw-test.jpg


--------------------------------------------------------------------------------
/face-alignment_projection/test/facealignment_test.py:
--------------------------------------------------------------------------------
 1 | import unittest
 2 | import face_alignment
 3 | 
 4 | 
 5 | class Tester(unittest.TestCase):
 6 |     def test_predict_points(self):
 7 |         fa = face_alignment.FaceAlignment(face_alignment.LandmarksType._3D, device='cpu')
 8 |         fa.get_landmarks('test/assets/aflw-test.jpg')
 9 | 
10 | if __name__ == '__main__':
11 |     unittest.main()
12 | 


--------------------------------------------------------------------------------
/face-alignment_projection/test/smoke_test.py:
--------------------------------------------------------------------------------
1 | import torch
2 | import face_alignment
3 | 


--------------------------------------------------------------------------------
/face-alignment_projection/test/test_utils.py:
--------------------------------------------------------------------------------
 1 | import unittest
 2 | from face_alignment.utils import *
 3 | import numpy as np
 4 | import torch
 5 | 
 6 | 
 7 | class Tester(unittest.TestCase):
 8 |     def test_flip_is_label(self):
 9 |         # Generate the points
10 |         heatmaps = torch.from_numpy(np.random.randint(1, high=250, size=(68, 64, 64)).astype('float32'))
11 | 
12 |         flipped_heatmaps = flip(flip(heatmaps.clone(), is_label=True), is_label=True)
13 | 
14 |         assert np.allclose(heatmaps.numpy(), flipped_heatmaps.numpy())
15 | 
16 |     def test_flip_is_image(self):
17 |         fake_image = torch.torch.rand(3, 256, 256)
18 |         fliped_fake_image = flip(flip(fake_image.clone()))
19 | 
20 |         assert np.allclose(fake_image.numpy(), fliped_fake_image.numpy())
21 | 
22 |     def test_getpreds(self):
23 |         pts = torch.from_numpy(np.random.randint(1, high=63, size=(68, 2)).astype('float32'))
24 | 
25 |         heatmaps = np.zeros((68, 256, 256))
26 |         for i in range(68):
27 |             if pts[i, 0] > 0:
28 |                 heatmaps[i] = draw_gaussian(heatmaps[i], pts[i], 2)
29 |         heatmaps = torch.from_numpy(np.expand_dims(heatmaps, axis=0))
30 | 
31 |         preds, _ = get_preds_fromhm(heatmaps)
32 | 
33 |         assert np.allclose(pts.numpy(), preds.numpy(), atol=5)
34 | 
35 | if __name__ == '__main__':
36 |     unittest.main()
37 | 


--------------------------------------------------------------------------------
/model/D.py:
--------------------------------------------------------------------------------
 1 | import torch
 2 | import torch.nn as nn
 3 | import torch.nn.functional as F
 4 | from torch.nn import init
 5 | from torch.nn import utils
 6 | 
 7 | from model.resblocks import Block
 8 | from model.resblocks import OptimizedBlock
 9 | 
10 | 
11 | class D(nn.Module):
12 | 
13 |     def __init__(self,input_nc=6 ,num_features=64, num_classes=0, activation=F.relu):
14 |         super(D, self).__init__()
15 |         self.num_features = num_features
16 |         self.num_classes = num_classes
17 |         self.activation = activation
18 |         if num_classes == None:
19 |             raise  ValueError('no given num_classes')
20 |         self.block1 = OptimizedBlock(input_nc, num_features)
21 |         self.block2 = Block(num_features, num_features * 2,
22 |                             activation=activation, downsample=True)
23 |         self.block3 = Block(num_features * 2, num_features * 4,
24 |                             activation=activation, downsample=True)
25 |         self.block4 = Block(num_features * 4, num_features * 8,
26 |                             activation=activation, downsample=True)
27 |         self.block5 = Block(num_features * 8, num_features * 8,
28 |                             activation=activation, downsample=True)  # 8
29 | 
30 | 
31 |         self.w_0 = nn.Parameter(torch.Tensor(1,num_features * 8))
32 |         nn.init.xavier_normal_(self.w_0.data)
33 |         self.b = nn.Parameter(torch.zeros(1))
34 | 
35 |         if num_classes > 0:
36 |             self.W = utils.spectral_norm(
37 |                 nn.Embedding(num_classes, num_features * 8))
38 | 
39 |     def _initialize(self):
40 |         optional_W = getattr(self, 'W', None)
41 |         if optional_W is not None:
42 |             init.xavier_uniform_(optional_W.weight.data)
43 | 
44 | 
45 |     def forward(self, x, y):  # 形如[1,23,45,..] 是视频序列，范围在0 到 num_class-1
46 |         # h = x
47 |         h1 = self.block1(x)
48 |         h2 = self.block2(h1)
49 |         h3 = self.block3(h2)
50 |         h4 = self.block4(h3)
51 |         h5 = self.block5(h4)
52 |         h5 = self.activation(h5)
53 |         # Global pooling
54 |         v_loss = torch.sum(h5, dim=(2, 3)) # B,C
55 |         v_loss = torch.sum(v_loss * (self.W(y) + self.w_0.squeeze()),dim=1) + self.b  # ([B])
56 |         v_loss = torch.mean(v_loss)
57 |         return [v_loss, h1, h2, h3, h4, h5,self.W(y)]
58 | 
59 | 
60 | if __name__=='__main__':
61 |     model = D(num_classes=100).cuda()
62 |     y = torch.ones([]).long().cuda()
63 |     data = torch.randn(1,6,256,256).cuda()
64 |     output = model(data,y)
65 |     v_loss  = output[0]
66 |     print(v_loss.size())  # torch.Size([])
67 |     w = output[-1]
68 |     print(w.size())       # torch.Size([8, 512])
69 | 


--------------------------------------------------------------------------------
/model/E.py:
--------------------------------------------------------------------------------
 1 | import torch
 2 | import torch.nn as nn
 3 | import torch.nn.functional as F
 4 | from torch.nn import init
 5 | from torch.nn import utils
 6 | 
 7 | from model.resblocks import Block
 8 | from model.resblocks import OptimizedBlock
 9 | 
10 | 
11 | class E(nn.Module):
12 | 
13 |     def __init__(self, num_features=64, activation=F.relu):
14 |         super(E, self).__init__()
15 |         self.num_features = num_features
16 |         self.activation = activation
17 | 
18 |         self.block1 = OptimizedBlock(6, num_features) # 128
19 |         self.block2 = Block(num_features, num_features * 2,
20 |                             activation=activation, downsample=True)  # 64
21 |         self.block3 = Block(num_features * 2, num_features * 4,
22 |                             activation=activation, downsample=True) # 32
23 |         self.block4 = Block(num_features * 4, num_features * 8,
24 |                             activation=activation, downsample=True) # 16
25 |         self.block5 = Block(num_features * 8, num_features * 8,
26 |                             activation=activation, downsample=True)  # 8
27 |         self.l6 = utils.spectral_norm(nn.Linear(num_features * 8, 256))
28 |         self.l7 = utils.spectral_norm(nn.Linear(num_features * 8, 256))
29 |         self._initialize()
30 | 
31 |     def _initialize(self):
32 |         init.xavier_uniform_(self.l6.weight.data)
33 |         optional_l_y = getattr(self, 'l_y', None)
34 |         if optional_l_y is not None:
35 |             init.xavier_uniform_(optional_l_y.weight.data)
36 | 
37 |     def forward(self, x,):
38 |         h = self.block1(x)
39 |         h = self.block2(h)
40 |         h = self.block3(h)
41 |         h = self.block4(h)
42 |         h = self.block5(h)
43 |         h = self.activation(h)
44 |         # Global pooling
45 |         h = torch.sum(h, dim=(2, 3))
46 |         h = self.activation(h)   # 512
47 |         e = torch.mean(h,dim=0,keepdim=True)
48 |         out1 = self.l6(h)  # B,256  P矩阵
49 |         size = out1.size()  # B,256
50 |         assert len(size)==2, 'size is not right'
51 |         mean = out1.view(size[0],size[1],1,1)  # 512维
52 |         mean = torch.mean(mean,dim=0,keepdim=True)
53 |         std = self.l7(h) # B,256
54 |         std = std.view(size[0],size[1],1,1)
55 |         std = torch.mean(std,dim=0,keepdim=True)
56 |         return e, mean, std
57 | 
58 | 
59 | if __name__=='__main__':
60 |     model = E().cuda()
61 |     data = torch.randn(8,6,256,256).cuda()
62 |     output = model(data)
63 |     e = output[0]
64 |     mean = output[1]
65 |     print(e.size())  # torch.Size([1, 512])
66 |     print(mean.size()) # torch.Size([1, 256, 1, 1])
67 | 


--------------------------------------------------------------------------------
/model/FaceSwapModel.py:
--------------------------------------------------------------------------------
 1 | from model.D import D
 2 | from model.G import G
 3 | from model.E import E
 4 | import numpy as np
 5 | import torch
 6 | from torch.nn import Module
 7 | from torch import nn
 8 | from model.loss import VGGLoss,VGGFaceLoss,CNTLoss,AdvLoss,MCHLoss,DLoss
 9 | # divide whole model into E and G&D
10 | class GDModel(Module):
11 |     def __init__(self,num_classes):
12 |         super(GDModel, self).__init__()
13 |         self.g = G(input_nc=3)
14 |         self.d = D(num_classes=num_classes)
15 |         self.cntloss = CNTLoss()
16 |         self.advloss = AdvLoss()
17 |         self.mchloss = MCHLoss()
18 |         self.dloss = DLoss()
19 | 
20 | 
21 |     def g_forward(self,landmark):
22 |         return self.g(landmark)
23 |     def d_forward(self,fake_img,real_image,y):
24 |         fake_info, real_info = self.d(fake_img,y), self.d(real_image,y)
25 |         return fake_info, real_info
26 |     def cal_cnt_loss(self,fake_image, real_image):
27 |         return self.cntloss(fake_image,real_image)
28 |     def cal_adv_loss(self,fake_info,real_info):  # include FM loss
29 |         fake_v_loss = fake_info[0]
30 |         fake_v_features = fake_info[1:6]
31 |         real_v_features = real_info[1:6]
32 |         return self.advloss(fake_v_features,real_v_features,fake_v_loss)
33 |     def cal_mch_loss(self,fake_info):
34 |         w = fake_info[6]
35 |         return self.mchloss(w,self.e)
36 |     def cal_d_loss(self,fake_info, real_info):
37 |         fake_v_loss = fake_info[0]
38 |         reak_v_loss = real_info[0]
39 |         return self.dloss(reak_v_loss,fake_v_loss)
40 |     def update_GDModel(self,mean_y,std_y,e):
41 |         self.e = e
42 |         self.g.update_adain(mean_y,std_y)
43 | 
44 |     def for_test_inference(self,landmark,y,x):
45 |         x_landmark = torch.cat((landmark,x),1)
46 |         fake_image = self.g_forward(landmark)   # from -1 to 1
47 |         fake_landmark = torch.cat((landmark,fake_image),1)
48 |         fake_info,real_info = self.d_forward(fake_landmark,x_landmark,y)
49 |         g_loss = self.cal_cnt_loss(fake_image,x) + self.cal_adv_loss(fake_info,real_info) +\
50 |             self.cal_mch_loss(fake_info)
51 |         d_loss = self.cal_d_loss(fake_info,real_info)
52 |         print(g_loss.size())
53 |         print(d_loss.size())
54 | 
55 |     def forward(self, landmark,y,x):
56 |         x_landmark = torch.cat((landmark,x),1)
57 |         fake_image = self.g_forward(landmark)   # from -1 to 1
58 |         fake_landmark = torch.cat((landmark,fake_image),1)
59 |         fake_info,real_info = self.d_forward(fake_landmark,x_landmark,y)
60 |         g_loss = self.cal_cnt_loss(fake_image,x) + self.cal_adv_loss(fake_info,real_info) +\
61 |             self.cal_mch_loss(fake_info)
62 |         d_loss = self.cal_d_loss(fake_info,real_info)
63 |         return fake_image,g_loss,d_loss
64 | 
65 | 
66 | if __name__=='__main__':
67 |     landmark = torch.randn(2, 3, 224, 224).cuda()
68 |     y = torch.LongTensor(np.random.randint(0, 50, size=[2])).cuda()
69 |     x = torch.randn(2, 3, 224, 224).cuda()
70 |     x_landmark = torch.cat((landmark, x), 1)
71 |     e_net = E().cuda()
72 |     e, mean_y, std_y = e_net(x_landmark)
73 |     model = GDModel(num_classes=30,).cuda()
74 |     model.update_GDModel(mean_y,std_y,e)
75 |     model.for_test_inference(landmark,y,x)
76 | 
77 | 
78 | 


--------------------------------------------------------------------------------
/model/G.py:
--------------------------------------------------------------------------------
  1 | import torch
  2 | from torch.nn.utils import spectral_norm
  3 | from torch.nn import Module
  4 | from torch import nn
  5 | from torch.nn import init
  6 | import math
  7 | 
  8 | class G(Module):
  9 |     def __init__(self,input_nc,):
 10 |         super(G,self).__init__()
 11 |         activation = nn.ReLU()
 12 | 
 13 |         model = [nn.ReflectionPad2d(4),
 14 |                  spectral_norm(nn.Conv2d(input_nc,128,9,1,bias=False)),
 15 |                  nn.InstanceNorm2d(128,affine=True),
 16 |                  activation]
 17 |         model = nn.Sequential(*model)
 18 | 
 19 |         # downsampling
 20 |         in_nc = 128
 21 |         out_nc = 2* in_nc    # 256, 256
 22 |         for i in range(2):
 23 |             model.add_module('res_down%d'% i,Residual_Block(in_nc,out_nc))
 24 |             in_nc = out_nc
 25 |         # residual same resolution
 26 |         for i in range(4):
 27 |             model.add_module('res_iden%d'% i, Residual_Ident(in_nc,256,))
 28 | 
 29 |         # upsampling
 30 |         for i in range(2):
 31 |             model.add_module('up%d'%i, Up(in_nc,in_nc))
 32 |             # in_nc = in_nc//2
 33 |         # reduce dimension from 64 to 3
 34 |         model.add_module('head',nn.Sequential(
 35 |             nn.ReflectionPad2d(1),
 36 |             spectral_norm(nn.Conv2d(256,128,3,1)),
 37 |             activation,
 38 | 
 39 |             nn.ReflectionPad2d(1),
 40 |             spectral_norm(nn.Conv2d(128, 64, 3, 1)),
 41 |             activation,
 42 | 
 43 |             nn.ReflectionPad2d(1),
 44 |             spectral_norm(nn.Conv2d(64, 3, 3, 1)),
 45 |             nn.Tanh()   # -1 ,1
 46 | 
 47 |         ))
 48 |         self.model = model
 49 |     def forward(self,x):
 50 |         x= self.model(x)
 51 |         # print('g mean : ',x.mean())
 52 |         return x
 53 |     def update_adain(self,mean_y,std_y):
 54 |         for m in self.modules():
 55 |             if isinstance(m,Adain):
 56 |                 m.update_mean_std(mean_y,std_y)
 57 | 
 58 | 
 59 | class Residual_Block(Module):
 60 |     def __init__(self,input_nc,output_nc):
 61 |         super(Residual_Block, self).__init__()
 62 |         activation = nn.ReLU(True)
 63 |         self.left = nn.Sequential(*[spectral_norm(nn.Conv2d(input_nc,output_nc,1,1,padding=0,bias=False)),
 64 |                                      nn.InstanceNorm2d(output_nc,affine=True),
 65 |                                      activation,
 66 |                                      nn.AvgPool2d(3, 2, padding=1)
 67 |                                      ])
 68 |         self.right = nn.Sequential(*[
 69 |             nn.ReflectionPad2d(1),
 70 |             spectral_norm(nn.Conv2d(input_nc,output_nc,3,1,padding=0,bias=False)),
 71 |             nn.InstanceNorm2d(output_nc,affine=True),
 72 |             activation,
 73 |             nn.ReflectionPad2d(1),
 74 |             spectral_norm(nn.Conv2d(output_nc, output_nc, 3, 1, padding=0, bias=False)),
 75 |             nn.InstanceNorm2d(output_nc,affine=True),
 76 |             activation,
 77 |             nn.AvgPool2d(3, 2, padding=1)
 78 |         ])
 79 | 
 80 |     def _initialize(self):
 81 |         init.xavier_uniform_(self.c1.weight.data, math.sqrt(2))
 82 |         for m in self.modules():
 83 |             if isinstance(m, nn.Conv2d):
 84 |                 n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
 85 |                 m.weight.data.normal_(0, math.sqrt(2. / n))
 86 |             elif isinstance(m, nn.InstanceNorm2d):
 87 |                 m.weight.data.fill_(1)
 88 |                 m.bias.data.zero_()
 89 |             else:
 90 |                 raise ValueError('No this layer init way')
 91 | 
 92 |     def forward(self, x):
 93 |         x1 = self.right(x)
 94 |         x2 = self.left(x)
 95 |         return x1+x2
 96 | 
 97 | class Residual_Ident(Module):
 98 |     def __init__(self,in_nc,out_nc,):
 99 |         super(Residual_Ident,self).__init__()
100 |         activation = nn.ReLU(True)
101 |         model = [nn.ReflectionPad2d(1),
102 |                  spectral_norm(nn.Conv2d(in_nc,out_nc,3,stride=1,padding=0,bias=False)),
103 |                  nn.InstanceNorm2d(out_nc,affine=False),
104 |                  Adain(),
105 |                  activation
106 |                  ]
107 |         self.model = nn.Sequential(*model)
108 |     def forward(self, x):
109 |         return self.model(x)
110 | 
111 | class Adain(Module):
112 |     def __init__(self,):
113 |         super(Adain,self).__init__()
114 | 
115 |     def forward(self, x):
116 |         size = x.size()
117 |         x = x*self.std_y.expand(size) + self.mean_y.expand(size)
118 |         return x
119 |     def update_mean_std(self,mean_y,std_y):   # be used before forward
120 |         self.mean_y = mean_y
121 |         self.std_y = std_y
122 | 
123 | class Up(Module):
124 |     def __init__(self,in_nc,out_nc):
125 |         super(Up, self).__init__()
126 |         activation = nn.ReLU(True)
127 |         model = [
128 |                 nn.Upsample(scale_factor=2,mode='bilinear'),
129 |                 nn.ReflectionPad2d(1),
130 |                 spectral_norm(nn.Conv2d(in_nc,out_nc,3,1,bias=False)),
131 |                 nn.InstanceNorm2d(out_nc,affine=False),
132 |                 Adain(),
133 |                 activation
134 |                  ]
135 |         self.model = nn.Sequential(*model)
136 |     def forward(self, x):
137 |         x = self.model(x)
138 |         return x
139 | 
140 | if __name__=='__main__':
141 |     std_y = torch.randn(2,256,1,1).cuda()
142 |     x = torch.randn(2,6,224,224).cuda()
143 |     mean_y = torch.zeros(2,256,1,1).cuda()
144 | 
145 |     model = G(6).cuda()
146 |     model.update_adain(mean_y,std_y)
147 |     out = model(x)
148 |     print(out.size())   # torch.Size([2, 3, 224, 224])
149 | 
150 | 
151 | 
152 |     # for m in model.modules():
153 |     #     if isinstance(m,Adain):
154 |     #         m.update_mean_std(mean_y,std_y)
155 |     # for m in model.children():
156 |     #     for c in m.modules():
157 |     #         if isinstance(c,Adain):
158 |     #             # print(c.mean_y==mean_y)
159 |     #             # print(c.std_y==std_y)
160 |     #             print('*********') # 6次


--------------------------------------------------------------------------------
/model/__init__.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__init__.py


--------------------------------------------------------------------------------
/model/__pycache__/D.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__pycache__/D.cpython-36.pyc


--------------------------------------------------------------------------------
/model/__pycache__/D.cpython-37.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__pycache__/D.cpython-37.pyc


--------------------------------------------------------------------------------
/model/__pycache__/E.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__pycache__/E.cpython-36.pyc


--------------------------------------------------------------------------------
/model/__pycache__/E.cpython-37.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__pycache__/E.cpython-37.pyc


--------------------------------------------------------------------------------
/model/__pycache__/FaceSwapModel.cpython-37.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__pycache__/FaceSwapModel.cpython-37.pyc


--------------------------------------------------------------------------------
/model/__pycache__/G.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__pycache__/G.cpython-36.pyc


--------------------------------------------------------------------------------
/model/__pycache__/G.cpython-37.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__pycache__/G.cpython-37.pyc


--------------------------------------------------------------------------------
/model/__pycache__/__init__.cpython-37.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__pycache__/__init__.cpython-37.pyc


--------------------------------------------------------------------------------
/model/__pycache__/loss.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__pycache__/loss.cpython-36.pyc


--------------------------------------------------------------------------------
/model/__pycache__/loss.cpython-37.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__pycache__/loss.cpython-37.pyc


--------------------------------------------------------------------------------
/model/__pycache__/resblocks.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__pycache__/resblocks.cpython-36.pyc


--------------------------------------------------------------------------------
/model/__pycache__/resblocks.cpython-37.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__pycache__/resblocks.cpython-37.pyc


--------------------------------------------------------------------------------
/model/__pycache__/vgg_face.cpython-36.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__pycache__/vgg_face.cpython-36.pyc


--------------------------------------------------------------------------------
/model/__pycache__/vgg_face.cpython-37.pyc:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/model/__pycache__/vgg_face.cpython-37.pyc


--------------------------------------------------------------------------------
/model/loss.py:
--------------------------------------------------------------------------------
  1 | from torchvision import models
  2 | import torch
  3 | import torch.nn as nn
  4 | import functools
  5 | from torch.autograd import Variable
  6 | import numpy as np
  7 | from torchvision.transforms import transforms
  8 | from model.vgg_face import vgg_m_face_bn_dag
  9 | 
 10 | class Vgg19(torch.nn.Module):
 11 |     def __init__(self, requires_grad=False):
 12 |         super(Vgg19, self).__init__()
 13 |         vgg_pretrained_features = models.vgg19(pretrained=True).features
 14 |         self.slice1 = torch.nn.Sequential()
 15 |         self.slice2 = torch.nn.Sequential()
 16 |         self.slice3 = torch.nn.Sequential()
 17 |         self.slice4 = torch.nn.Sequential()
 18 |         self.slice5 = torch.nn.Sequential()
 19 |         for x in range(2):
 20 |             self.slice1.add_module(str(x), vgg_pretrained_features[x])
 21 |         for x in range(1, 7):
 22 |             self.slice2.add_module(str(x), vgg_pretrained_features[x])
 23 |         for x in range(7, 12):
 24 |             self.slice3.add_module(str(x), vgg_pretrained_features[x])
 25 |         for x in range(12, 21):
 26 |             self.slice4.add_module(str(x), vgg_pretrained_features[x])
 27 |         for x in range(21, 30):
 28 |             self.slice5.add_module(str(x), vgg_pretrained_features[x])
 29 | 
 30 |         if not requires_grad:
 31 |             for param in self.parameters():
 32 |                 param.requires_grad = False
 33 | 
 34 |     def forward(self, X):
 35 |         # self.zero_grad()
 36 |         h_relu1 = self.slice1(X)
 37 |         h_relu2 = self.slice2(h_relu1)
 38 |         h_relu3 = self.slice3(h_relu2)
 39 |         h_relu4 = self.slice4(h_relu3)
 40 |         h_relu5 = self.slice5(h_relu4)
 41 |         out = [h_relu1, h_relu2, h_relu3, h_relu4, h_relu5]
 42 |         return out
 43 | 
 44 | class VGGLoss(nn.Module):
 45 |     def __init__(self):
 46 |         super(VGGLoss,self).__init__()
 47 |         self.vgg = Vgg19()
 48 |         self.criterion = nn.L1Loss(size_average=True)
 49 |         self.weight = [0.01] *5
 50 |     def forward(self, x,y):
 51 |         x_vgg, y_vgg = self.vgg(x),self.vgg(y)
 52 |         loss = 0
 53 |         for i in range(len(x_vgg)):
 54 |             loss = self.weight[i]/float(len(x_vgg)) * self.criterion(x_vgg[i],y_vgg[i].detach()) + loss
 55 |         return loss
 56 | class VGGFaceLoss(nn.Module):
 57 |     def __init__(self):
 58 |         super(VGGFaceLoss,self).__init__()
 59 |         self.vggface = vgg_m_face_bn_dag('./pretrained/vgg_m_face_bn_dag.pth')
 60 |         # self.vggface = vgg_m_face_bn_dag(None)
 61 |         self.criterion = nn.L1Loss(size_average=True)
 62 |         self.weights = [0.002] * 5
 63 |     def forward(self, x,y):
 64 |         x_vggface, y_vggface = self.vggface(x), self.vggface(y)
 65 |         loss = 0
 66 |         for i in range(len(x_vggface)):
 67 |             loss += self.weights[i]/5.0 * self.criterion(x_vggface[i],y_vggface[i].detach())
 68 |         return loss
 69 | 
 70 | class CNTLoss(nn.Module):
 71 |     def __init__(self):
 72 |         super(CNTLoss, self).__init__()
 73 |         self.vggloss = VGGLoss()
 74 |         self.vggfaceloss = VGGFaceLoss()
 75 |         self.vgg_preprocessed = transforms.Normalize([0.485,0.456,0.406],
 76 |                                                      [0.229,0.224,0.225])
 77 |     def forward(self,x,y):
 78 |         x = (x+1)*127.5
 79 |         y = (y+1)*127.5
 80 |         x1 = self.vgg_preprocessed(x.squeeze(0)/255.0).unsqueeze(0)
 81 |         y1 = self.vgg_preprocessed(y.squeeze(0)/255.0).unsqueeze(0)
 82 |         loss = 0.5 * (self.vggfaceloss(x,y) + self.vggloss(x1,y1))
 83 |         return loss
 84 | 
 85 | 
 86 | class AdvLoss(nn.Module):
 87 |     def __init__(self,):
 88 |         super(AdvLoss,self).__init__()
 89 |         self.criterion = nn.L1Loss(size_average=True)
 90 | 
 91 |     def forward(self, fake_feature,real_feature,v_loss):
 92 |         fm_loss = 0
 93 |         feat_weights = 10.0 / len(fake_feature)
 94 |         for i in range(len(fake_feature)-1):
 95 |             fm_loss += feat_weights * self.criterion(fake_feature[i],real_feature[i].detach())
 96 |         return  -v_loss + fm_loss  # D_loss必须是标量才行
 97 | 
 98 | 
 99 | class MCHLoss(nn.Module):
100 |     def __init__(self):
101 |         super(MCHLoss,self).__init__()
102 |         self.criterion = nn.L1Loss(size_average=True)
103 | 
104 |     def forward(self, w,e):
105 |         return 80 * self.criterion(w, e.detach())
106 | 
107 | class DLoss(nn.Module):
108 |     def __init__(self):
109 |         super(DLoss,self).__init__()
110 |     def forward(self, real_vloss, fake_vloss):
111 |         d_loss = torch.mean(torch.relu(1. - real_vloss)) +\
112 |             torch.mean(torch.relu(1. + fake_vloss))
113 |         return d_loss
114 | 


--------------------------------------------------------------------------------
/model/resblocks.py:
--------------------------------------------------------------------------------
 1 | import math
 2 | 
 3 | import torch
 4 | import torch.nn as nn
 5 | import torch.nn.functional as F
 6 | from torch.nn import init
 7 | from torch.nn import utils
 8 | 
 9 | 
10 | class Block(nn.Module):
11 | 
12 |     def __init__(self, in_ch, out_ch, h_ch=None, ksize=3, pad=1,
13 |                  activation=F.relu, downsample=False):
14 |         super(Block, self).__init__()
15 | 
16 |         self.activation = activation
17 |         self.downsample = downsample
18 | 
19 |         self.learnable_sc = (in_ch != out_ch) or downsample
20 |         if h_ch is None:
21 |             h_ch = in_ch
22 |         else:
23 |             h_ch = out_ch
24 | 
25 |         self.c1 = utils.spectral_norm(nn.Conv2d(in_ch, h_ch, ksize, 1, pad))
26 |         self.c2 = utils.spectral_norm(nn.Conv2d(h_ch, out_ch, ksize, 1, pad))
27 |         if self.learnable_sc:
28 |             self.c_sc = utils.spectral_norm(nn.Conv2d(in_ch, out_ch, 1, 1, 0))
29 | 
30 |         self._initialize()
31 | 
32 |     def _initialize(self):
33 |         init.xavier_uniform_(self.c1.weight.data, math.sqrt(2))
34 |         init.xavier_uniform_(self.c2.weight.data, math.sqrt(2))
35 |         if self.learnable_sc:
36 |             init.xavier_uniform_(self.c_sc.weight.data)
37 | 
38 |     def forward(self, x):
39 |         return self.shortcut(x) + self.residual(x)
40 | 
41 |     def shortcut(self, x):
42 |         if self.learnable_sc:
43 |             x = self.c_sc(x)
44 |         if self.downsample:
45 |             return F.avg_pool2d(x, 2)
46 |         return x
47 | 
48 |     def residual(self, x):
49 |         h = self.c1(self.activation(x))
50 |         h = self.c2(self.activation(h))
51 |         if self.downsample:
52 |             h = F.avg_pool2d(h, 2)
53 |         return h
54 | 
55 | 
56 | class OptimizedBlock(nn.Module):
57 | 
58 |     def __init__(self, in_ch, out_ch, ksize=3, pad=1, activation=F.relu):
59 |         super(OptimizedBlock, self).__init__()
60 |         self.activation = activation
61 | 
62 |         self.c1 = utils.spectral_norm(nn.Conv2d(in_ch, out_ch, ksize, 1, pad))
63 |         self.c2 = utils.spectral_norm(nn.Conv2d(out_ch, out_ch, ksize, 1, pad))
64 |         self.c_sc = utils.spectral_norm(nn.Conv2d(in_ch, out_ch, 1, 1, 0))
65 | 
66 |         self._initialize()
67 | 
68 |     def _initialize(self):
69 |         init.xavier_uniform_(self.c1.weight.data, math.sqrt(2))
70 |         init.xavier_uniform_(self.c2.weight.data, math.sqrt(2))
71 |         init.xavier_uniform_(self.c_sc.weight.data)
72 | 
73 |     def forward(self, x):
74 |         return self.shortcut(x) + self.residual(x)
75 | 
76 |     def shortcut(self, x):
77 |         return self.c_sc(F.avg_pool2d(x, 2))
78 | 
79 |     def residual(self, x):
80 |         h = self.activation(self.c1(x))
81 |         return F.avg_pool2d(self.c2(h), 2)
82 | 


--------------------------------------------------------------------------------
/model/vgg_face.py:
--------------------------------------------------------------------------------
 1 | 
 2 | import torch
 3 | import torch.nn as nn
 4 | 
 5 | 
 6 | class Vgg_m_face_bn_dag(nn.Module):
 7 | 
 8 |     def __init__(self):
 9 |         super(Vgg_m_face_bn_dag, self).__init__()
10 |         self.meta = {'mean': [131.45376586914062, 103.98748016357422, 91.46234893798828],
11 |                      'std': [1, 1, 1],
12 |                      'imageSize': [224, 224, 3]}
13 |         self.mean = torch.FloatTensor(self.meta['mean']).unsqueeze(0).unsqueeze(2).unsqueeze(3).cuda()
14 |         self.resize = torch.nn.UpsamplingBilinear2d(size=(224,224))
15 |         self.conv1 = nn.Conv2d(3, 96, kernel_size=[7, 7], stride=(2, 2))
16 |         self.bn49 = nn.BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
17 |         self.relu1 = nn.ReLU()
18 |         self.pool1 = nn.MaxPool2d(kernel_size=[3, 3], stride=[2, 2], padding=0, dilation=1, ceil_mode=False)
19 |         self.conv2 = nn.Conv2d(96, 256, kernel_size=[5, 5], stride=(2, 2), padding=(1, 1))
20 |         self.bn50 = nn.BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
21 |         self.relu2 = nn.ReLU()
22 |         self.pool2 = nn.MaxPool2d(kernel_size=[3, 3], stride=[2, 2], padding=(0, 0), dilation=1, ceil_mode=True)
23 |         self.conv3 = nn.Conv2d(256, 512, kernel_size=[3, 3], stride=(1, 1), padding=(1, 1))
24 |         self.bn51 = nn.BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
25 |         self.relu3 = nn.ReLU()
26 |         self.conv4 = nn.Conv2d(512, 512, kernel_size=[3, 3], stride=(1, 1), padding=(1, 1))
27 |         self.bn52 = nn.BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
28 |         self.relu4 = nn.ReLU()
29 |         self.conv5 = nn.Conv2d(512, 512, kernel_size=[3, 3], stride=(1, 1), padding=(1, 1))
30 |         self.bn53 = nn.BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
31 |         self.relu5 = nn.ReLU()
32 |         self.pool5 = nn.MaxPool2d(kernel_size=[3, 3], stride=[2, 2], padding=0, dilation=1, ceil_mode=False)
33 |         self.fc6 = nn.Conv2d(512, 4096, kernel_size=[6, 6], stride=(1, 1))
34 |         self.bn54 = nn.BatchNorm2d(4096, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
35 |         self.relu6 = nn.ReLU()
36 |         self.fc7 = nn.Conv2d(4096, 4096, kernel_size=[1, 1], stride=(1, 1))
37 |         self.bn55 = nn.BatchNorm2d(4096, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
38 |         self.relu7 = nn.ReLU()
39 |         self.fc8 = nn.Linear(in_features=4096, out_features=2622, bias=True)
40 | 
41 |         for param in self.parameters():
42 |             param.requires_grad = False
43 |         # for m in self.modules():
44 |         #     if isinstance(m,nn.BatchNorm2d):
45 |         #         m = m.eval()
46 | 
47 |     def forward(self, x0):
48 |         # self.zero_grad()
49 |         x0 = self.resize(x0)
50 |         x0 = x0 - self.mean.expand(x0.size())
51 |         x1 = self.conv1(x0)
52 |         x2 = self.bn49(x1)
53 |         x3 = self.relu1(x2)
54 |         x4 = self.pool1(x3)
55 |         x5 = self.conv2(x4)
56 |         x6 = self.bn50(x5)
57 |         x7 = self.relu2(x6)
58 |         x8 = self.pool2(x7)
59 |         x9 = self.conv3(x8)
60 |         x10 = self.bn51(x9)
61 |         x11 = self.relu3(x10)
62 |         x12 = self.conv4(x11)
63 |         x13 = self.bn52(x12)
64 |         x14 = self.relu4(x13)
65 |         x15 = self.conv5(x14)
66 |         x16 = self.bn53(x15)
67 |         x17 = self.relu5(x16)
68 |         x18 = self.pool5(x17)
69 |         x19 = self.fc6(x18)
70 |         x20 = self.bn54(x19)
71 |         x21 = self.relu6(x20)
72 |         x22 = self.fc7(x21)
73 |         x23 = self.bn55(x22)
74 |         x24_preflatten = self.relu7(x23)
75 |         x24 = x24_preflatten.view(x24_preflatten.size(0), -1)
76 |         x25 = self.fc8(x24)
77 |         return [x1, x6, x11, x18, x25]
78 | 
79 | def vgg_m_face_bn_dag(weights_path=None, **kwargs):
80 |     """
81 |     load imported model instance
82 | 
83 |     Args:
84 |         weights_path (str): If set, loads model weights from the given path
85 |     """
86 |     model = Vgg_m_face_bn_dag()
87 |     if weights_path:
88 |         state_dict = torch.load(weights_path)
89 |         model.load_state_dict(state_dict)
90 |     return model


--------------------------------------------------------------------------------
/train.py:
--------------------------------------------------------------------------------
  1 | import torch
  2 | import numpy as np
  3 | import os
  4 | from model.E import E
  5 | from model.FaceSwapModel import GDModel
  6 | from dataset.reader import get_loader
  7 | from torch.optim import Adam
  8 | from utils import update_learning_rate
  9 | from torch import nn
 10 | import visdom
 11 | import cv2
 12 | 
 13 | torch.manual_seed(0)
 14 | torch.cuda.manual_seed_all(0)
 15 | 
 16 | torch.backends.cudnn.benchmark = True
 17 | torch.backends.cudnn.deterministic = True
 18 | # vis = visdom.Visdom()
 19 | 
 20 | clip_txt = './dataset/video_clips_path.txt'
 21 | batchsize = 1
 22 | num_workers = 4
 23 | epoches = 60
 24 | loader, num_classes = get_loader(clip_txt,batchsize,num_workers)
 25 | E = E().cuda()
 26 | E.train()
 27 | model = GDModel(num_classes).cuda()
 28 | model.train()
 29 | for m in model.cntloss.vggfaceloss.vggface.modules():
 30 |     if isinstance(m,nn.BatchNorm2d):
 31 |         m.eval()
 32 | 
 33 | # 涉及到batchnorm 在training阶段，要计算batch的均值和标准差，使得batchsize不能为1，但是eval的话，又没法计算梯度
 34 | # model.cntloss.vggfaceloss.vggface.eval()
 35 | # model.cntloss.vggloss.vgg.eval()
 36 | #
 37 | 
 38 | # mutilple GPUs
 39 | # E = torch.nn.DataParallel(E,)
 40 | # model = torch.nn.DataParallel(model)
 41 | 
 42 | 
 43 | # define optim
 44 | g_e_parameters = list(E.parameters())
 45 | g_e_parameters += list(model.g.parameters())
 46 | 
 47 | lr = 2e-4
 48 | g_e_optim = Adam(g_e_parameters,lr = lr ,betas=(0.5,0.999) )
 49 | d_optim = Adam(model.d.parameters(),lr = 5 * lr,betas=(0.5,0.999))
 50 | 
 51 | global_step = 0
 52 | for e in range(epoches):
 53 |     current_clip_number = 0
 54 |     print('current epoch is %d'%epoches)
 55 |     for d in loader:
 56 |         print('current_clip_number is %d' % current_clip_number)
 57 |         # calculate e, std_y, mean_y for adaptive instance norm
 58 |         data_for_e = d['imgs_e']
 59 |         data_for_e = torch.cat(data_for_e,0).cuda()
 60 |         landmark_for_e = d['landmarks_e']
 61 |         landmark_for_e = torch.cat(landmark_for_e,0).cuda()
 62 |         batch_data = d['imgs_training']
 63 |         batch_data = torch.cat(batch_data,0)
 64 |         batch_landmark = d['landmarks_training']
 65 |         batch_landmark = torch.cat(batch_landmark,0)
 66 | 
 67 |         # e,mean_y,std_y = E(torch.cat((data_for_e,landmark_for_e),1))
 68 |         # model.update_GDModel(mean_y,std_y,e)
 69 | 
 70 |         # print(data_for_e.size())
 71 |         # print(landmark_for_e.size())
 72 |         for b,l in zip(batch_data,batch_landmark):  # b and l are 3-dim tensors
 73 |             
 74 |             e, mean_y, std_y = E(torch.cat((data_for_e, landmark_for_e), 1))
 75 |             model.update_GDModel(mean_y, std_y, e)
 76 | 
 77 |             global_step += 1
 78 |             b = b.unsqueeze(0).cuda()
 79 |             l = l.unsqueeze(0).cuda()
 80 |             y = torch.tensor(current_clip_number).long().cuda()
 81 |             fake_img, g_loss, d_loss = model(l,y,b)
 82 | 
 83 |             model.cntloss.vggfaceloss.vggface.zero_grad()
 84 |             model.cntloss.vggloss.vgg.zero_grad()
 85 |             g_e_optim.zero_grad()
 86 |             g_loss.backward(retain_graph=True)
 87 |             g_e_optim.step()
 88 | 
 89 | 
 90 |             d_optim.zero_grad()
 91 |             d_loss.backward()
 92 |             d_optim.step()
 93 | 
 94 |             if global_step%100==0:
 95 |                 fake_img = np.transpose(np.uint8((fake_img.cpu().data.numpy()[0]/2.0 + 0.5)*255),[1,2,0])
 96 |                 b_ = np.transpose(np.uint8((b.cpu().data.numpy()[0]/2.0 + 0.5)*255),[1,2,0])
 97 |                 l_ = np.transpose(np.uint8((l.cpu().data.numpy()[0]/2.0 + 0.5)*255),[1,2,0])
 98 |                 # temp = np.stack((fake_img,b_,l_))  # 3, 3 ,256,256
 99 |                 temp = np.concatenate((fake_img[:,:,::-1],b_[:,:,::-1],l_[:,:,::-1]),axis=1)
100 |                 cv2.imwrite('./training_visual/temp_fake_gt_landmark_%d.jpg'%global_step,temp)
101 |             # vis.images(temp,nrow=1,win='temp_results')
102 | 
103 |             print('***************')
104 |         current_clip_number += 1
105 |         if global_step % 50 == 0:
106 |             saved = {'e':E.state_dict(),
107 |                      'g_d': model.state_dict()}
108 |             torch.save(saved,'./saved_models/e_g_d%d.pth'%global_step)
109 |     if (e+1)%10 ==0:
110 |         lr = lr/2.0
111 |         update_learning_rate(g_e_optim,lr)
112 |         update_learning_rate(d_optim,5*lr)
113 | 


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_100.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_100.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_1000.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_1000.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_1100.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_1100.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_1200.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_1200.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_1300.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_1300.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_1400.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_1400.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_1500.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_1500.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_1600.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_1600.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_1700.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_1700.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_1800.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_1800.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_1900.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_1900.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_200.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_200.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_2000.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_2000.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_2100.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_2100.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_2200.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_2200.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_2300.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_2300.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_300.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_300.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_400.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_400.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_500.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_500.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_600.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_600.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_700.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_700.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_800.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_800.jpg


--------------------------------------------------------------------------------
/training_visual/temp_fake_gt_landmark_900.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/shoutOutYangJie/Few-Shot-Adversarial-Learning-for-face-swap/927b10bcf066507caba470bdf66bf543fcb2520b/training_visual/temp_fake_gt_landmark_900.jpg


--------------------------------------------------------------------------------
/utils.py:
--------------------------------------------------------------------------------
 1 | from torch.optim import Adam
 2 | from torchvision.models import vgg16
 3 | 
 4 | def update_learning_rate(optim,lr):
 5 |     for param_group in optim.param_groups:
 6 |         param_group['lr'] = lr
 7 | 
 8 | 
 9 | if __name__=='__main__':
10 |     model = vgg16()
11 |     optim = Adam(model.parameters(),lr=3)
12 |     print(optim.param_groups[0]['lr'])
13 |     update_learning_rate(optim,0.001)
14 |     print(optim.param_groups[0]['lr'])


--------------------------------------------------------------------------------