Genebody-Benchmarks

This repository contains the training and evaluation code for NeuralBody, NeuralVolumes, NeuralHumanRendering, NeuralTexture, A-NeRF and IBRNet to perform novel-view synthesis on Genebody dataset. Following benchmark tables are also shown in the paper.

The code for each method is on the branches of this repository. To re-implement the results on GeneBody, please download the pretrained models in the Model Zoo first, and prepare the environment and dataset based on the README.md on each branch.

News

[29/04/22]: First version of benchmarks released, containing 5 case-specific methods and 1 generalizable methods.

Benchmarks

Case-specific Methods on Genebody

Model	PSNR	SSIM	LPIPS	ckpts
NV	19.86	0.774	0.267	ckpts
NHR	20.05	0.800	0.155	ckpts
NT	21.68	0.881	0.152	ckpts
NB	20.73	0.878	0.231	ckpts
A-Nerf	15.57	0.508	0.242	ckpts

(see detail why A-Nerf's performance is counterproductive in issue)

Generalizable Methods on Genebody

Model	PSNR	SSIM	LPIPS	ckpts
PixelNeRF (Our implemetation coming soon)	24.15	0.903	0.122
IBRNet	23.61	0.836	0.177	ckpts

Citation

@article{cheng2022generalizable,
    title={Generalizable Neural Performer: Learning Robust Radiance Fields for Human Novel View Synthesis},
    author={Cheng, Wei and Xu, Su and Piao, Jingtan and Qian, Chen and Wu, Wayne and Lin, Kwan-Yee and Li, Hongsheng},
    journal={arXiv preprint arXiv:2204.11798},
    year={2022}
}