TODO

1. Create 'data' directory and subdirectory for target video ex) 'data/studio1'.

2. Create subdirectory for source video(or audio) and textgrid in 'data' directory ex) 'data/studio_2_0_test'

3. Create 'clip' directory in 'data/studio1' directory, and put video clips ex) 'data/studio1/clip/studio_1_0.mp4'.

4. Create 'textgrid' directory in 'data/studio1' directory, and put textgrid files ex) 'data/studio1/textgrid/studio_1_0.TextGrid'

5. Put test video(or audio) and textgrid file in subdirectory created in step 2. ex) 'data/studio_2_0_test/studio_2_0.mp4', 'data/studio_2_0_test/studio_2_0.TextGrid'

6. In https://drive.google.com/drive/u/0/folders/11LuLtRMU-f_AVO0hOI031bXSx19g7IEe , download all files and

   • put s3fd.pth in 'audiodvp_utils/face_detection/detection/sfd' directory
   • create 'weights' directory and put 'resnet50_ft_weight.pkl'

7. Follow 'scripts/merge.sh' to create merged video and flist.txt file

8. Follow 'scripts/phoneme_match.sh' for lip sync generation

TODO

1. Create 'data' directory and subdirectory for target video and source video ex) 'data/target', 'data/source.

2. Put video for target and source directory ex) 'data/target/target.mp4', 'data/source/source.mp4.

3. In https://drive.google.com/drive/u/0/folders/11LuLtRMU-f_AVO0hOI031bXSx19g7IEe , download all files and

   • put s3fd.pth in 'audiodvp_utils/face_detection/detection/sfd' directory
   • create 'weights' directory and put 'resnet50_ft_weight.pkl'
   • put 'vgg19_conv.pth' in 'vendor/neural_face_renderer/models/' directory

4. Follow Usage step 1~6 in 'https://github.com/xinwen-cs/AudioDVP'

5. Follow 'scripts/face_reenactment.sh' for face reenactment