Flow-guided One-shot Talking Face Generation with a High-resolution Audio-visual Dataset paper

./HDTF_dataset consists of youtube video url, time stamps of talking face and facial region in the video. xx_video_url.txt:

format:     video name | video youtube url

xx_annotion_time.txt:

format:    video name | time stamps of clip1 | time stamps of clip2 | time stamps of clip3....

xx_crop_wh.txt:

format:    video name+clip index | min_width | width |  min_height | height

If you use HDTF dataset, pls

1. Download videos from xx_video_url.txt with you-get tool or youtube-dl tool. (pls download the highest definition version: 1080P or 720P). Transform video into .mp4 format. You'd better transform interlaced video to progressive video as well.

2. Split long original video into appropriate talking head clips with time stamps in xx_annotion_time.txt. Name the splitted clip as video name_clip index.mp4. For example, split the video Radio11.mp4 00:30-01:00 01:30-02:30 into Radio11_0.mp4 and Radio11_1.mp4 .

3. Crop the facial region with fixed window size in xx_crop_wh.txt and resize the video into 512 x 512 resolution.

if you use HDTF, pls reference

@inproceedings{zhang2021flow,
  title={Flow-Guided One-Shot Talking Face Generation With a High-Resolution Audio-Visual Dataset},
  author={Zhang, Zhimeng and Li, Lincheng and Ding, Yu and Fan, Changjie},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages={3661--3670},
  year={2021}
}