Reproduce “Synthesizing Dexterous Robot Demonstrations from Human Videos via Virtual Try-On” 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
Seungyong Lee, Suhyeok Jang, Minyoung Kim, Jinwoo Shin
RoboTryOn converts human manipulation videos into robot-looking videos, then infers robot actions from them to create policy-training data. It translates frames into robot keyframes, synthesizes motion guided by hand-skeleton motion, and uses an inverse dynamics model (IDM) for action labels. Across OpenArm and RB-Y1, adding conversions raises out-of-domain scores from 2.1% to 60.0% across seven tasks without task-specific robot demonstrations.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.