VolumetricSMPL:
A Neural Volumetric Body Model for
Efficient Interactions, Contacts, and Collisions



Overview

VolumetricSimple is a lightweight and user-friendly add-on module for SMPL(-X)-based body models, enabling seamless volumetric extension. With just a single line of code, users can extend SMPL models with volumetric functionalities. After completing the forward pass, they gain access to key volumetric functionalities, including SDF queries, self-intersection loss, and collision penalties. This implementation maintains full compatibility with existing SMPL-based reconstruction and perception applications.




Check out our paper for more results and comparisons.

BibTex

@inproceedings{ICCV25:VolumetricSMPL,
   title={{VolumetricSMPL}: A Neural Volumetric Body Model for Efficient Interactions, Contacts, and Collisions},
   author={Mihajlovic, Marko and Zhang, Siwei and Li, Gen and Zhao, Kaifeng and M{\"u}ller, Lea and Tang, Siyu},
   booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
   year={2025}
}


Contact

For questions, please contact Marko Mihajlovic or raise an issue on GitHub.