README.md

HumanNeRF

A pytorch implementation of HumanNeRF (CVPR2022), as described in HumanNeRF : Efficiently Generated Human Radiance Field from Sparse Inputs.

HumanNeRF : Efficiently Generated Human Radiance Field from Sparse Inputs
Fuqiang Zhao, Wei Yang, Jiakai Zhang, Pei Lin, Yingliang Zhang, Jingyi Yu, Lan Xu
Project page / Paper / Video / Data

Installation

Tested on Ubuntu 20.04 + Pytorch 1.10.1 + RTX3090

Install environment:

$ conda create -n humannerf python=3.8
$ conda activate humannerf
$ pip install torch==1.10.1+cu113 torchvision==0.11.2+cu113 torchaudio==0.10.1+cu113 -f https://download.pytorch.org/whl/cu113/torch_stable.html
$ pip install imageio pillow scikit-image opencv-python configargparse lpips kornia warmup_scheduler matplotlib test-tube imageio-ffmpeg

Training

Please see each subsection for training or inference on different componenments. Available training datasets:

Dataset

Our collected multi-view datasets are avaliable at Google Driver

Training nerf model

Run

$ cd tools
$ python train_net.py configs/config.yml --gpu 0

Testing

Testing nerf model

Run

$ cd tools
$ python render.py ${DATA_DIR} --output ${OUTPU_DIR} --render nerf --gpu 0

Testing texture blending model

Run

$ cd tools
$ python render.py ${DATA_DIR} --output ${OUTPU_DIR} --render blending --gpu 0

License and Citation

@InProceedings{Zhao_2022_CVPR,
    author    = {Zhao, Fuqiang and Yang, Wei and Zhang, Jiakai and Lin, Pei and Zhang, Yingliang and Yu, Jingyi and Xu, Lan},
    title     = {HumanNeRF: Efficiently Generated Human Radiance Field From Sparse Inputs},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    month     = {June},
    year      = {2022},
    pages     = {7743-7753}
}