redshoga / head-changer

Detect the face direction from the video and display the image in front of the face

Head Changer

Based on

https://github.com/yinguobing/head-pose-estimation

How to use

Sample Data

Please refer to /sample-data

Create head changed serial number frame images

sudo docker build -t redshoga/head-changer .
sudo docker run -it -v $(pwd):/share redshoga/head-changer python /share/changer.py --video /share/sample-data/video1.mp4 --image /share/sample-data/4.png --output-dir /share/output

Transfer frame images to video

ffmpeg -r 30 -i ./output/%d.png -vcodec libx264 -pix_fmt yuv420p -r 30 ./sample-data/out.mp4

TODO

• Add example gif video to README.md
• Refactor scripts
• Add transfer frame images to video using FFmpeg in script