# Phy-SIC **Repository Path**: to4698_admin/Phy-SIC ## Basic Information - **Project Name**: Phy-SIC - **Description**: [SIGGGRAPHASIA2025] PhySIC: Physically Plausible 3D Human-Scene Interaction and Contact from a Single Image - **Primary Language**: Unknown - **License**: MIT - **Default Branch**: main - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 0 - **Created**: 2026-11-03 - **Last Updated**: 2026-11-03 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README

PhySIC: Physically Plausible 3D Human-Scene Interaction and Contact from a Single Image

[![arXiv](https://img.shields.io/badge/Arxiv-2510.11649-b31b1b.svg?logo=arXiv)](http://arxiv.org/abs/2510.11649) [![Home Page](https://img.shields.io/badge/Project-Website-C27185.svg)](https://yuxuan-xue.com/physic/) [Pradyumna Yalandur Muralidhar](https://pradyumanym.github.io/)★, [Yuxuan Xue](https://yuxuan-xue.com/)★†, [Xianghui Xie](https://virtualhumans.mpi-inf.mpg.de/people/Xie.html), Margaret Kostyrko, [Gerard Pons-Moll](https://virtualhumans.mpi-inf.mpg.de/)
TL;DR: Human-Scene Interaction and Contact from a Single Image with Physical Plausibility.

https://github.com/user-attachments/assets/86105b9b-ff49-41de-8fd8-dbf25e3e0b26 ## Getting Started ### Installation Please follow the instructions in [INSTALL.md](INSTALL.md) to set up the environment. ### Downloading Data Please run the following script to download the required data: ```bash bash fetch_data.sh ``` This requires access to [SMPL](https://smpl.is.tue.mpg.de/), [SMPL-X](https://smpl-x.is.tue.mpg.de/), [AGORA](https://agora.is.tue.mpg.de/), and [CameraHMR](https://camerahmr.is.tue.mpg.de/). Please enter the credentials when prompted. ### Running the Code To run the code, please copy images to the `images/` directory. Then, execute the following command: ```bash python run_optimizer.py ``` This will save the results in the `outputs/` directory. ## Roadmap - [x] **Demo code release** - [ ] **Evaluation code** ## Acknowledgements This code is built on top of many great open-source projects. We would like to thank the authors of the following repositories: - [OmniEraser](https://github.com/PRIS-CV/Omnieraser) for the image inpainting code. - [MoGe](https://github.com/microsoft/MoGe) for affine-invariant depth estimation. - [CameraHMR](https://github.com/pixelite1201/CameraHMR/) for the initial human mesh estimation. - [WiLoR](https://github.com/rolpotamias/WiLoR)/[WiLoR-mini](https://github.com/warmshao/WiLoR-mini) for initial hand pose estimation. - [DECO](https://github.com/sha2nkt/deco) and [PROX](https://github.com/mohamedhassanmus/prox) for contact estimation and static contacts. - [DepthPro](https://github.com/apple/ml-depth-pro) for metric depth estimation. - [HSfM](https://github.com/hongsukchoi/HSfM_RELEASE) and [ViTPose](https://github.com/ViTAE-Transformer/ViTPose) for the ViTPose inference code and models. - [LangSAM](https://github.com/luca-medeiros/lang-segment-anything), [GroundingDINO](https://github.com/IDEA-Research/GroundingDINO), and [Segment-Anything](https://github.com/facebookresearch/segment-anything) for the segmentation code and models. - [SMPLFitter](https://github.com/isarandi/smplfitter)/[NLF](https://github.com/isarandi/nlf) for the SMPL-to-SMPL-X converter. and many others. ## Citation If you find our work useful, please cite: ```bibtex @inproceedings{ym2025physic, author = {Yalandur Muralidhar, Pradyumna and Xue, Yuxuan and Xie, Xianghui and Kostyrko, Margaret and Pons-Moll, Gerard}, title = {PhySIC: Physically Plausible 3D Human-Scene Interaction and Contact from a Single Image}, journal = {SIGGRAPH Asia 2025 Conference Papers}, year = {2025}, } ```