Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces AURORA, an active 3D reconstruction framework that enhances in-hand object observation by integrating online object-centric reconstruction with a novel reorientation strategy. By employing Ray-GPIS to estimate direction-wise reconstruction uncertainty and selecting next-best-view targets based on an uncertainty-novelty objective, AURORA significantly improves the quality of reconstruction and efficiency of information acquisition. Experimental results show that AURORA outperforms traditional non-active rotation strategies and existing active view-planning methods in various performance metrics, demonstrating its robustness against occlusions and pose errors.
AURORA achieves superior 3D reconstruction quality and efficiency by actively targeting under-observed regions through uncertainty-driven reorientation.
Observing objects grasped by a robot hand is challenging due to severe visual occlusions. Although in-hand manipulation can expose hidden surfaces, existing approaches often rely on predefined or open-loop reorientation strategies that do not explicitly target under-observed regions. We propose AURORA, an active 3D reconstruction framework that closes the loop between online object-centric reconstruction and in-hand reorientation. At its core, Ray-GPIS estimates direction-wise reconstruction uncertainty along candidate viewing rays and selects next-best-view targets using an uncertainty--novelty objective, which are realized through an axis-conditioned in-hand rotation policy. The resulting RGB-D observations are fused incrementally using CAD-free 6D pose tracking and lightweight geometric reconstruction. Experiments demonstrate that AURORA improves reconstruction quality and information-acquisition efficiency over non-active rotation strategies, while Ray-GPIS also outperforms active view-planning baselines in reconstruction performance, action-ranking quality, and planning efficiency. Targeted ablations further validate its robustness to hand occlusion and pose errors. The project webpage is available at https://aurorahand.github.io/