This project is about face-swapping processing in video. Inspirations are iperov/DeepFaceLive and haofanwang/inswapper.
- Face detection & face swapping: deepinsight/insightface
- The model used for swapping is ezioruan/inswapper_128.onnx
- Upsampling of swapped images: sczhou/CodeFormer
Suppose you want to swap your face with the image below,
Source by Pexels.com
InsightFace recognizes your face and replaces yours to the input image.
The swapped image is a little rough and CodeFormers upsamples face area.
Some codes are borrowed from sczhou/CodeFormer
basicsr/
facelib/
weights/
FaceRestorationHelperWrapper.py
largely stems frominference_codeformer.py
pip install opencv-python torch torchvision insightface