- Notifications
You must be signed in to change notification settings - Fork8
[CVPR 2023] UV Volumes for Real-time Rendering of Editable Free-view Human Performance
License
fanegg/UV-Volumes
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
Project Page |Paper |Latest arXiv |Supplementary
UV Volumes for Real-time Rendering of Editable Free-view Human Performance
Yue Chen*,Xuan Wang*,Xingyu Chen,Qi Zhang,Xiaoyu Li,Yu Guo†,Jue Wang,Fei Wang
(* equal contribution，† corresponding author)
CVPR 2023
This repository is an official implementation ofUV-Volumes usingpytorch.
Please seeINSTALL.md for manual installation.
Please seeINSTALL.md to download the dataset.
Take the training onsequence 313 as an example.
python3 train_net.py --cfg_file configs/zju_mocap_exp/313.yaml exp_name zju313 resume False output_depth TrueYou can monitor the training process by Tensorboard.
tensorboard --logdir data/record/UVvolume_ZJUTake the test onsequence 313 as an example.
python3 run.py --type evaluate --cfg_file configs/zju_mocap_exp/313.yaml exp_name zju313 use_lpips True test.frame_sampler_interval 1 use_nb_mask_at_box True save_img True T_threshold 0.75Please seeINSTALL.md to download and process the dataset.
Take the training on171204_pose4_sample6 as an example.
python3 train_net.py --cfg_file configs/cmu_exp/p4s6.yaml exp_name p4s6 resume False output_depth TrueYou can monitor the training process by Tensorboard.
tensorboard --logdir data/record/UVvolume_CMUTake the test on171204_pose4_sample6 as an example.
python3 run.py --type evaluate --cfg_file configs/cmu_exp/p4s6.yaml exp_name p4s6 use_lpips True test.frame_sampler_interval 1 use_nb_mask_at_box True save_img TrueIf you find this code useful for your research, please use the following BibTeX entry.
@inproceedings{chen2023uv,title={UV Volumes for real-time rendering of editable free-view human performance},author={Chen, Yue and Wang, Xuan and Chen, Xingyu and Zhang, Qi and Li, Xiaoyu and Guo, Yu and Wang, Jue and Wang, Fei},booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},pages={16621--16631},year={2023}}
Our code is based on the awesome pytorch implementation ofNeuralBody. We appreciate all the contributors.
About
[CVPR 2023] UV Volumes for Real-time Rendering of Editable Free-view Human Performance