Reproduce “EgoHand: Monocular 3D Hand Reconstruction from Egocentric Video” 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
Chao-Yeh Chen, Tiffany Chen, Lin Duan, Tianyi Hu, Yu-Hsiang Huang, Umar Iqbal, Jiefeng Li, Edy Lim and 7 more
EgoHand reconstructs 3D hand trajectories from first-person video using camera-aware crop context and predicted 2D joints to place reconstructed hands. It is trained on a 14-million-image corpus and evaluated zero-shot on H2O, HOT3D, and ARCTIC. On H2O, it achieves 13.12 mm mean MPJPE versus WiLoR’s 17.77 mm, while WiLoR performs better on PA-MPJPE and 2D EPE; evaluations use ground-truth hand boxes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.