Reproduce “SelfPose: 3D Egocentric Pose Estimation from a Headset Mounted Camera” 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
Denis Tome, Thiemo Alldieck, Patrick Peluse, Gerard Pons-Moll, Lourdes Agapito, Hernan Badino, Fernando De la Torre
SelfPose estimates full-body 3D joint positions and rotations from a downward fisheye headset image. Its model detects joint heatmaps, then reconstructs pose with auxiliary rotation and heatmap-reconstruction targets intended to retain uncertainty information. On Mo2Cap2’s outdoor test set, it records 54.61 mm versus 80.64 mm for Mo2Cap2, after training only on provided synthetic data; extreme occlusion can still cause failures.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.