ObjSplat: Geometry-Aware Gaussian Surfels for Active Object Reconstruction

2026-01-11

arXiv.org

AI Summary

This work proposes an active reconstruction framework based on a unified Gaussian surfel representation to address the geometric incompleteness and inefficiency caused by conventional viewpoint selection strategies in autonomous high-fidelity object reconstruction. The approach introduces a geometry-aware viewpoint evaluation mechanism that explicitly models back-facing visibility and occlusion-aware multi-view co-visibility, coupled with a multi-step optimal path planner (NBP) that balances information gain against motion cost. Efficient exploration is achieved through dynamic construction of a spatial graph. Experiments demonstrate that the method generates physically consistent, high-fidelity complete models within minutes in both simulated environments and real-world cultural artifact scanning, significantly outperforming existing approaches in terms of surface completeness, reconstruction quality, scanning time, and path length.

Technology Category

• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Computer Vision: 3D Computer Vision
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Autonomous high-fidelity object reconstruction is fundamental for creating digital assets and bridging the simulation-to-reality gap in robotics. We present ObjSplat, an active reconstruction framework that leverages Gaussian surfels as a unified representation to progressively reconstruct unknown objects with both photorealistic appearance and accurate geometry. Addressing the limitations of conventional opacity or depth-based cues, we introduce a geometry-aware viewpoint evaluation pipeline that explicitly models back-face visibility and occlusion-aware multi-view covisibility, reliably identifying under-reconstructed regions even on geometrically complex objects. Furthermore, to overcome the limitations of greedy planning strategies, ObjSplat employs a next-best-path (NBP) planner that performs multi-step lookahead on a dynamically constructed spatial graph. By jointly optimizing information gain and movement cost, this planner generates globally efficient trajectories. Extensive experiments in simulation and on real-world cultural artifacts demonstrate that ObjSplat produces physically consistent models within minutes, achieving superior reconstruction fidelity and surface completeness while significantly reducing scan time and path length compared to state-of-the-art approaches. Project page: https://li-yuetao.github.io/ObjSplat-page/ .

Problem

• object reconstruction
• autonomous scanning
• geometry-aware
• high-fidelity
• active reconstruction

Innovation

• Gaussian surfels
• geometry-aware viewpoint evaluation
• next-best-path planning
• active object reconstruction
• multi-view covisibility