English

A Finite Difference Approximation of Second Order Regularization of Neural-SDFs

Graphics 2025-11-13 v1 Computer Vision and Pattern Recognition Machine Learning

Abstract

We introduce a finite-difference framework for curvature regularization in neural signed distance field (SDF) learning. Existing approaches enforce curvature priors using full Hessian information obtained via second-order automatic differentiation, which is accurate but computationally expensive. Others reduced this overhead by avoiding explicit Hessian assembly, but still required higher-order differentiation. In contrast, our method replaces these operations with lightweight finite-difference stencils that approximate second derivatives using the well known Taylor expansion with a truncation error of O(h^2), and can serve as drop-in replacements for Gaussian curvature and rank-deficiency losses. Experiments demonstrate that our finite-difference variants achieve reconstruction fidelity comparable to their automatic-differentiation counterparts, while reducing GPU memory usage and training time by up to a factor of two. Additional tests on sparse, incomplete, and non-CAD data confirm that the proposed formulation is robust and general, offering an efficient and scalable alternative for curvature-aware SDF learning.

Keywords

Cite

@article{arxiv.2511.08980,
  title  = {A Finite Difference Approximation of Second Order Regularization of Neural-SDFs},
  author = {Haotian Yin and Aleksander Plocharski and Michal Jan Wlodarczyk and Przemyslaw Musialski},
  journal= {arXiv preprint arXiv:2511.08980},
  year   = {2025}
}

Comments

SIGGRAPH Asia Technical Communications, 6 pages, 6 figures, preprint

R2 v1 2026-07-01T07:33:23.372Z