Search papers, labs, and topics across Lattice.
This paper introduces OccamView, an innovative object-conditioned view-selection framework designed for active 3D Gaussian reconstruction under strict frame budgets. By leveraging an online object memory and occlusion-aware proxy-coverage scores, OccamView strategically prioritizes views that enhance the reconstruction of partially observed objects rather than merely focusing on global scene coverage. Experimental results on Replica and Matterport3D datasets reveal that OccamView significantly improves completion rates, particularly when frame budgets are constrained, showcasing the effectiveness of integrating object-specific cues with traditional geometry-driven methods.
Object-conditioned cues can dramatically enhance 3D reconstruction quality, especially under tight frame budgets, outperforming traditional geometry-focused strategies.
Active 3D Gaussian reconstruction fundamentally relies on selecting informative next-best views under limited sensing budgets. Existing active 3DGS methods primarily plan viewpoints according to geometric information gain, treating object-induced hidden regions in the same manner as general unexplored space. Under tight frame budgets, such geometry-driven strategies may prioritize global scene coverage while leaving partially observed objects incompletely reconstructed. To address this limitation, we propose OccamView, an object-conditioned view-selection framework for frame-budgeted active 3D Gaussian reconstruction. Rather than predicting unseen object geometry or performing shape completion, OccamView maintains an online object memory from open-vocabulary detections grounded in measured RGB-D observations and represents unresolved local occupancy around detected objects as conservative hidden-region proxies. Candidate viewpoints are then evaluated using an occlusion-aware proxy-coverage score. Furthermore, we introduce a Geo-Floor mechanism that restricts object-conditioned re-ranking to geometrically competitive candidates, allowing object-conditioned cues to guide complementary observations while preserving the geometry-driven exploration behavior of the underlying planner. Experiments on Replica and Matterport3D under a unified frame-budgeted protocol show that OccamView consistently reduces Completion and improves Completion Ratio across five frame budgets, with particularly pronounced gains under limited frame budgets. These results demonstrate that lightweight object-conditioned cues effectively complement geometry-driven active view planning.
