NerfCap: Human Performance Capture With Dynamic Neural Radiance Fields

Project Page | Video | Paper

Kangkan Wang, Sida Peng, Xiaowei Zhou, Jian Yang, Guofeng Zhang

Official implementation of Human Performance Capture With Dynamic Neural Radiance Fields.

Questions and discussions are welcomed! Feel free to contact Kangkan Wang via wangkangkan@njust.edu.cn

Installation

Set up python environment

conda create -n nerfcap python=3.7
conda activate nerfcap

pip install torch==1.6.0+cu102 -f https://download.pytorch.org/whl/torch_stable.html
pip install -r requirements.txt

Set up dataset

Download DeepCap dataset here.

Run the code on DeepCap Dataset

Test

Download the corresponding pretrained model and put it to $ROOT/data/trained_model/if_nerf/magdalena/latest.pth
Test and visualization
- Visualize all frames at test views python run.py --type visualize --cfg_file configs/magdalena/magdalena.yaml exp_name magdalena
- Simultaneously extract mesh at each frame python run.py --type visualize --cfg_file configs/magdalena/magdalena.yaml exp_name magdalena vis_mesh True
The result are located at $ROOT/data/result/if_nerf/magdalena

Train

Train python train_net.py --cfg_file configs/magdalena/magdalena.yaml exp_name magdalena resume False
Tensorboard tensorboard --logdir data/record/if_nerf

Citation

If you find this code useful for your research, please use the following BibTeX entry.

@ARTICLE{NerfCap_TVCG22,
title={NerfCap: Human Performance Capture With Dynamic Neural Radiance Fields}, 
author={Wang, Kangkan and Peng, Sida and Zhou, Xiaowei and Yang, Jian and Zhang, Guofeng},  
journal={IEEE Transactions on Visualization and Computer Graphics},   
year={2022},  
pages={1-13},  
doi={10.1109/TVCG.2022.3202503}}

wangkangkan/nerfcap