Collision-Aware and Observation-Aligned Object-Centric Scene Reconstruction from Point Cloud

2026-09-29

AI Summary

This study addresses the challenges of scale ambiguity, insufficient geometric constraints, and object collisions in object completion for point cloud scene reconstruction by proposing the COOL framework. This method conditions a generative model on instance and background point clouds to achieve observation-aligned, object-centric scene reconstruction. A core innovation lies in the introduction of an explicit collision loss function, combined with joint optimization and resampling strategies, which effectively suppresses inter-object penetration during generation. Experimental results demonstrate that COOL significantly improves scene fidelity and geometric alignment on datasets such as 3D-Front, while exhibiting strong robustness to mask errors.

Abstract

Object-centric scene reconstruction requires completing partial object observations while preserving metric alignment and avoiding collisions with the surrounding. Existing generation-based methods are often image-conditioned and suffer from scale ambiguity and insufficient geometric constraints. We propose COOL, a framework for COllision-aware and Observation-aLigned reconstruction. Based on an object generation model, COOL conditions the generation on instance and background point clouds. Instance geometry anchors generation in scene coordinates, while background geometry provides local context for scene-consistent completion. We further introduce an explicit collision loss and use joint optimization and resampling to reduce collisions during inference. Experiments on 3D-Front and Scan2CAD demonstrate strong scene-level fidelity, observation alignment, and collision reduction. Moreover, additional studies validate its robustness to mask errors and its applicability to real-world scene replicas.

Problem

• object-centric scene reconstruction
• point cloud completion
• scale ambiguity
• collision avoidance
• geometric constraints

Innovation

• Object-Centric Scene Reconstruction
• Point Cloud Conditioning
• Collision-Aware Generation
• Observation Alignment
• Joint Optimization