Skip to content

[NeurIPS 2024] Generalizable and Animatable Gaussian Head Avatar

License

Notifications You must be signed in to change notification settings

xg-chu/GAGAvatar_track

Repository files navigation

⚡️ Tracking Framework for GAGAvatar ⚡️

     

🚀 Track video 🚀
drawing
🚅 Track image 🚅
drawing

Description

GAGAvatar Track is a monocular face tracker built on FLAME. It provides FLAME parameters (including eyeball pose) and camera parameters, along with the bounding box and landmarks used during optimization.

Installation

Build environment

This environment is a sub-environment of GAGAvatar. You can skip this step if you have already built GAGAvatar.

conda env create -f environment.yml
conda activate GAGAvatar_track

Prepare resources

Prepare resources with bash ./build_resources.sh.

Resources Link

The models and resources are available at https://huggingface.co/xg-chu/GAGAvatar_track.

Fast start

It takes longer to track the first frame.

Track on video(s):

python track_video.py -v ./demos/obama.mp4

Track on image(s):

python track_image.py -i ./demos/monroe.jpg

Track all images in a LMDB dataset:

python track_lmdb.py -l ./demos/vfhq_demo

Citation

If you find our work useful in your research, please consider citing:

@inproceedings{
    chu2024gagavatar,
    title={Generalizable and Animatable Gaussian Head Avatar},
    author={Xuangeng Chu and Tatsuya Harada},
    booktitle={The Thirty-eighth Annual Conference on Neural Information Processing Systems},
    year={2024},
    url={https://openreview.net/forum?id=gVM2AZ5xA6}
}

Acknowledgements

Some part of our work is built based on FLAME, StyleMatte, EMICA and VGGHead. The GAGAvatar Logo is designed by Caihong Ning. We thank you for sharing their wonderful code and their wonderful work.

About

[NeurIPS 2024] Generalizable and Animatable Gaussian Head Avatar

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published