发表机构
State Key Laboratory of Virtual Reality Technology and Systems, Beihang University; School of Computer Science and Engineering, Beihang University; Zhongguancun Laboratory（北京航空航天大学虚拟现实技术与系统国家重点实验室; 北京航空航天大学计算机科学与工程学院; 中关村实验室）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
GSPotential是面向稀疏视图3D Gaussian Splatting的框架，通过相机势场解决稀疏视图下的过拟合与几何伪影，采用虚拟相机采样与高斯更新策略提升重建保真度且训练效率具竞争力。
AI 中文摘要
三维高斯溅射（3D Gaussian Splatting）在光真实感渲染领域取得了显著成功，但由于光度监督的固有缺陷，其在稀疏视图场景中存在严重过拟合和几何伪影问题。近期进展尝试通过融合深度、点云或扩散模型等外部先验来正则化优化，但这些方法通常忽略了视图空间中监督的非均匀分布，导致先验使用的特异性有限且控制方式原始。本文提出GSPotential框架，利用相机势场量化视图空间的监督不平衡问题，核心思路是识别光度约束最欠缺的监督谷，并通过该势场从两个互补方面指导重建：一是设计概率球形采样策略，在低势区域放置有信息的虚拟相机，这些视图的点云渲染结果提供针对性几何指导；二是该势场为弱覆盖空间区域的保守高斯更新提供方向覆盖提示。大量实验表明，GSPotential在保持竞争力训练效率的同时，实现了高重建保真度。
英文摘要
3D Gaussian Splatting has achieved remarkable success in photorealistic rendering, yet it suffers from severe overfitting and geometric artifacts in sparse-view scenarios due to the inherent deficiency of photometric supervision. Recent advances have attempted to regularize optimization by incorporating external priors, such as depth, point clouds, or diffusion models. However, these methods typically overlook the non-uniform distribution of supervision across the viewing space, resulting in limited specificity in prior use and primitive control. In this paper, we propose GSPotential, a framework that quantifies view-space supervision imbalance using a Camera Potential Field. Our key insight is to identify supervision valleys where photometric constraints are most deficient, and use the potential field to guide reconstruction from two complementary aspects. First, we devise a probabilistic spherical sampling strategy that places informative virtual cameras in low-potential regions. Point-cloud renderings from these views then provide targeted geometric guidance. Second, the same field provides a directional coverage cue for conservative Gaussian updates in weakly covered spatial sectors. Extensive experiments demonstrate that GSPotential achieves high reconstruction fidelity while maintaining competitive training efficiency.
Comments10 pages, 8 figures, 5 tables. Accepted to Pacific Graphics 2026 Conference Track
