Face animation using Ai Thin-Plate Spline Motion Model

Thin-Plate Spline Motion Mode is an Ai model to add movement to a still image based on the driving video. The paper is here.

1. Prepare a short video mp4 with a face, resize it to 512×512. Prepare a face photo (or drawing) with the same size. The size and position of faces in the video and image are around the same.
2. Go to Google colab. Click the first triangle button to run git command.
Click “Run anyway” to close pop up. Wait for green check mark.
3. Open the folder icon at left, you see a folder Thin-Plate-Spline-Motion-Model is created. Run the next block to change directory.
4. Run the third block to retrieve checkpoints vox.pth.tar. If you see error and no file is showing in checkpoints folder, you have to manually get vox.path.tar and put in checkpoints folder.
5. Open assets folder, delete the original source.png and driving.mp4. Rename your own video file and headshot file to the same names respectively. Drag them in assets folder.
6. Run the next two blocks starting with “import torch” and “try”. Wait until your files are showing below.
7. Run the next block to load checkpoints.
8. Run the last block to make face animation using Ai model. Wait for video generated and download it.