Reproduce “PhysTacGen: Physics-Aware Visual-Tactile Sensor Image Generation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Guo Tang, Yongtao Wang
PhysTacGen generates optical tactile sensor images from RGB images by combining material-property descriptions with relative-depth cues. On a curated SSVTP test split, its GTPO prompting achieves SSIM 0.964, compared with 0.948 for supervised fine-tuning, while LPIPS is 0.559 versus 0.570. These results measure image similarity, not physical correctness, force prediction, or robot-ready performance.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.