Reproduce “RoPT: Robust 6DoF Object Pose Tracking in 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
Yongqing Liang, Yuzhong Chen, Sihang Li, Junsuk Kim, Chase Kidder, Hui-ping Wang, Yunzhi Xu, Raihan Haque and 2 more
RoPT is an RGB-D robot-object tracking system that combines mask-derived translation, graph-based motion prediction, alternative pose hypotheses, and failure-triggered reacquisition with FoundationPose refinement. On YCBInEOAT at 16× frame-rate downsampling, it reports ADD and ADD-S AUC scores of 90.85 and 94.55, versus 67.37 and 79.55 for FoundationPose++. The system assumes an exact CAD model and can fail under prolonged occlusion.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.