WildHead.DatProc

Refactor

filter

size_filter: To filter image size.

blur_filter: To filter blurry image.

detector

head_detector: To detect head.

visualize

vis_2d: To visualize 2d images.

Cropper

front_view_cropper: To crop front view images.

Misc

Data Processing for SphereHead

Notes

Thanks to PanoHead code for data processing. This repository is a fork of the original code with modification.

• 3DDFA_V2 git branch tag is 1b6c67601abffc1e9f248b291708aef0e43b55ae.

Configure Conda Environment

conda create -n datproc python=3.8
conda activate datproc
conda install pytorch==1.12.1 torchvision==0.13.1 torchaudio==0.12.1 cudatoolkit=11.3 -c pytorch
conda install cudnn==8.2.1
pip install onnxruntime-gpu==1.14.1
pip install face-alignment
pip install opencv-python==4.5.4.58
pip install matplotlib imageio imageio-ffmpeg pyyaml tqdm argparse cython scikit-image scipy gradio
pip install dlib

Demo

python dlib_kps.py
python recrop_images.py -i examples/test/data.pkl -j dataset.json -o examples/test/crop_samples/img

Citation

If you find our repo helpful, please cite our paper using the following bib:

@misc{li2024spherehead,
    title={SphereHead: Stable 3D Full-head Synthesis with Spherical Tri-plane Representation},
    author={Heyuan Li and Ce Chen and Tianhao Shi and Yuda Qiu and Sizhe An and Guanying Chen and Xiaoguang Han},
    year={2024},
    eprint={2404.05680},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}