Reproduce “PhySIC: Physically Plausible 3D Human-Scene Interaction and Contact from a Single Image” on OpenResearch

About this paper

Pradyumna Yalandur Muralidhar, Yuxuan Xue, Xianghui Xie, Margaret Kostyrko, Gerard Pons-Moll

Researchers from the University of Tübingen developed PhySIC, a framework reconstructing metric-scale 3D human and scene geometries, alongside dense vertex-level contact maps, from a single monocular image. The method reduces human pose error (PA-MPJPE) to 41.99mm on PROX and 46.50mm on RICH-100, while achieving contact F1-scores of 0.511 and 0.428 respectively, demonstrating robust physical plausibility and efficient reconstruction.

Key insights

• Jointly reasoning about human and scene geometry within a unified optimization framework is crucial for resolving monocular ambiguities and achieving physically consistent 3D reconstructions.
• Selectively applying interpenetration losses only to visible human vertices, while robustly regularizing occluded parts, is critical for maintaining local pose accuracy and preventing unrealistic deformation.
• Combining metric-scale depth estimation with detailed relative point maps, and explicitly fitting a ground plane, provides robust initialization and essential context for physically plausible human-scene interaction.

Summary by alphaXiv.