Confidence-Based Mesh Extraction from 3D Gaussians

2026-03-25

AI Summary

This work addresses the challenge of surface ambiguity in extracting high-quality meshes from 3D Gaussian representations within unbounded scenes exhibiting strong view-dependent appearance. To mitigate the adverse impact of appearance variations on geometry reconstruction, the authors introduce a self-supervised confidence mechanism into 3D Gaussian splatting for the first time. This mechanism employs learnable confidence scores to dynamically weight photometric and geometric supervision. Furthermore, the approach incorporates color and normal variance regularization alongside a decoupled D-SSIM appearance model. Notably, the method achieves state-of-the-art mesh extraction quality without relying on multi-view fusion, iterative optimization, or large models, thereby maintaining computational efficiency while significantly improving geometric fidelity.

Technology Category

• Computer Vision: 3D Computer Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Recently, 3D Gaussian Splatting (3DGS) greatly accelerated mesh extraction from posed images due to its explicit representation and fast software rasterization. While the addition of geometric losses and other priors has improved the accuracy of extracted surfaces, mesh extraction remains difficult in scenes with abundant view-dependent effects. To resolve the resulting ambiguities, prior works rely on multi-view techniques, iterative mesh extraction, or large pre-trained models, sacrificing the inherent efficiency of 3DGS. In this work, we present a simple and efficient alternative by introducing a self-supervised confidence framework to 3DGS: within this framework, learnable confidence values dynamically balance photometric and geometric supervision. Extending our confidence-driven formulation, we introduce losses which penalize per-primitive color and normal variance and demonstrate their benefits to surface extraction. Finally, we complement the above with an improved appearance model, by decoupling the individual terms of the D-SSIM loss. Our final approach delivers state-of-the-art results for unbounded meshes while remaining highly efficient.

Problem

• mesh extraction
• 3D Gaussian Splatting
• view-dependent effects
• geometric ambiguity
• surface reconstruction

Innovation

• confidence-based
• 3D Gaussian Splatting
• mesh extraction
• self-supervised
• view-dependent effects