用新型多项式核提升多视角3D重建的细节精度
High-Fidelity Surface Splatting-Based 3D Reconstruction from Multi-View Images
- 设计局部支持的多项式核替代指数核，更好控制高频细节
- 结合拉普拉斯滤波的随机正则化，稳定优化并保留精细结构
- 端到端重建效果优于现有方法，几何更准、图像更锐
多视角网格重建在计算机图形与视觉中仍是核心挑战，尤其在稀疏观测下恢复高频几何信息。当前方法如3D高斯点阵（3DGS）和神经辐射场（NeRF）依赖后处理提取网格，限制了几何与外观的联合优化。隐式移动最小二乘法（IMLS）可直接将点云转为有符号距离场与纹理场，支持端到端重建与渲染。但现有IMLS采用指数核，在高频细节恢复上表现不佳。本文提出一种紧凑的多项式核，具有局部支持与更高灵活性，能更好调控频率内容，提升几何保真度。为进一步增强细粒度结构，引入基于拉普拉斯滤波的随机正则化。实验表明，该方法在表面重建与渲染上达到领先性能，从多视角数据中生成更精确几何与更清晰视觉效果。
原文摘要 · Abstract (English)
Multi-view mesh reconstruction remains a core challenge in computer graphics and vision, especially for recovering high-frequency geometry from sparse observations. Recent methods such as 3D Gaussian Splatting (3DGS) and Neural Radiance Fields (NeRF) rely on post-processing for mesh extraction, thereby limiting joint optimization of geometry and appearance. Implicit Moving Least Squares (IMLS) instead enables direct conversion of point clouds into signed distance and texture fields, supporting end-to-end reconstruction and rendering. However, existing IMLS formulations use exponential kernels that struggle with high-frequency detail. We introduce a compact polynomial kernel with local support and greater flexibility, allowing better control over frequency content and improved geometric fidelity. To further enhance fine details, we incorporate stochastic regularization with Laplacian filtering. Together, these improve the preservation of high-frequency structure while maintaining stable optimization. Experiments show state-of-the-art performance in both surface reconstruction and rendering, yielding more accurate geometry and sharper visuals from multi-view data.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。