Reproduce “RoboMoCap: A Large-Scale Whole-Body Motion Capture Dataset with Egocentric Vision for Robot Learning” 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
Mayur Deshmukh, Aaron Tung, Junyi Li, Nikash Bhardwaj, Ademi Adeniji, Vincent Liu, Michael Cho, Jiafei Duan
RoboMoCap pairs first-person camera views with synchronized optical full-body motion capture to measure locomotion, including feet often outside view. The dataset contains over 200 hours from 14 participants. Median head-travel disagreement is 92.4° during lateral/backward locomotion versus 9.5° walking forward. Existing reconstruction methods are benchmarked on foot contact and placement, but the study does not show improved robot performance.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.