Search papers, labs, and topics across Lattice.
MAPRPose is a two-stage 6D object pose estimation framework that uses mask-aware correspondences for pose proposal and amodal-driven ROI prediction for refinement. The Mask-Aware Pose Proposal (MAPP) stage lifts 2D correspondences into 3D to generate pose hypotheses, while the refinement stage uses an Amodal Mask Prediction and ROI Re-Alignment (AMPR) module to handle occlusion. Evaluated on the BOP benchmark, MAPRPose achieves state-of-the-art Average Recall (AR) of 76.5% and a 43x speedup in multi-object inference compared to FoundationPose.
Achieve state-of-the-art 6D object pose estimation with 43x speedup by combining mask-aware pose proposals with amodal-driven refinement.
6D object pose estimation in cluttered scenes remains challenging due to severe occlusion and sensor noise. We propose MAPRPose, a two-stage framework that leverages mask-aware correspondences for pose proposal and amodal-driven Region-of-Interest (ROI) prediction for robust refinement. In the Mask-Aware Pose Proposal (MAPP) stage, we lift 2D correspondences into 3D space to establish reliable keypoint matches and generate geometrically consistent pose hypotheses based on correspondence-level scoring, from which the top-$K$ candidates are selected. In the refinement stage, we introduce a tensorized render-and-compare pipeline integrated with an Amodal Mask Prediction and ROI Re-Alignment (AMPR) module. By reconstructing complete object geometry and dynamically adjusting the ROI, AMPR mitigates localization errors and spatial misalignment under heavy occlusion. Furthermore, our GPU-accelerated RGB-XYZ reprojection enables simultaneous refinement of all $N \times B$ pose hypotheses in a single forward pass. Evaluated on the BOP benchmark, MAPRPose achieves a state-of-the-art Average Recall (AR) of 76.5%, outperforming FoundationPose by 3.1% AR while delivering a 43x speedup in multi-object inference.
