README.md

March 12, 2025 ยท View on GitHub

MEAT: Multiview Diffusion Model for Human Generation on Megapixels with Mesh Attention

TL;DR

With mesh attention designed for efficient cross-view feature fusion, MEAT is the first human
multiview diffusion model that can generate dense, view-consistent multiview images at a resolution of 1024x1024.

Project Page

Teaser

News

[03/2025] Our paper has been released to arxiv.

[03/2025] Paper and Code coming soon!

[02/2025] MEAT is accepted to CVPR 2025 :fire:

Citation

If you find our work useful for your research, please consider citing our paper:

@InProceedings{wang2025meat,
    title = {{MEAT}: Multiview Diffusion Model for Human Generation on Megapixels with Mesh Attention},
    author = {Wang, Yuhan and Hong, Fangzhou and Yang, Shuai and Jiang, Liming and Wu, Wayne and Loy, Chen Change},
    booktitle = {CVPR},
    year = {2025},
}