Reproduce “EgoMo3R: Joint Egocentric Motion and Scene Reconstruction” 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
Chaitanya Patel, Yusuke Kato, Rin Takahira, Kazuki Kozuka, Juan Carlos Niebles, Ehsan Adeli
EgoMo3R jointly reconstructs egocentric camera paths, scenes, and body motion from video. It alternates visual geometry and a learned motion prior so scene and movement constrain each other. On EgoExo4D, metric-scale mean Accuracy and Completion distances are 0.357 and 0.187, versus ViPE’s 0.905 and 0.819, supporting better scene reconstruction, not universal gains in camera-pose metrics.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.