#273 · Primary category: Computer Vision

ECON

3d-reconstruction avatar-generator computer-graphics computer-vision digital-twins metaverse normal-maps pifu pifuhd smpl-body smplx virtual-humans

[CVPR'23, Highlight] ECON: Explicit Clothed humans Optimized via Normal integration

Project last updated:09/17/24

GitHub Stars

1.2K

Forks

116

Contributors

11

License

Other

Why we included this project

For teams building virtual avatars or digital humans, ECON is a solid academic starting point: it turns a single color photo into a high-fidelity clothed 3D body, even when the subject wears loose clothing or strikes an unusual pose. The method blends implicit and explicit surface representations to get around the failure cases that trip up simpler approaches. The repo includes inference code, pretrained models, and a documented pipeline from normal prediction to final mesh, so you can adapt it rather than rebuild everything from zero. Just note the research-only license if commercial use is on the table. For prototyping and study, it's one of the more complete and accessible baselines around.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category