Summary
This paper introduces AmbiSuR, a framework for photometric ambiguity-robust 3D surface reconstruction built on Gaussian Splatting, authored by Jiahe Li, Jiawei Zhang, Xiao Bai, Jin Zheng, Xiaohan Yu, Lin Gu, and Gim Hee Lee (arXiv:2605.12494). While differentiable rendering has advanced surface reconstruction, pervasive photometric ambiguities remain a bottleneck. The authors revisit the foundations of Gaussian Splatting and identify two built-in primitive-wise ambiguities in the representation, while also discovering an intrinsic potential for ambiguity self-indication. Based on these insights, AmbiSuR first introduces photometric disambiguation, which constrains ill-posed geometry solutions to form definite surfaces, and then proposes an ambiguity indication module that identifies under-constrained reconstructions and guides their correction. Extensive experiments show the method outperforms existing surface reconstruction approaches across challenging scenarios and offers broad compatibility. This forum post shares the paper's abstract, arXiv link, and metadata for the computer vision research community.
Paper Overview
Field: Computer Vision (CV)
Authors: Jiahe Li, Jiawei Zhang, Xiao Bai, Jin Zheng, Xiaohan Yu, Lin Gu, Gim Hee Lee
Published: 2026-05-12
arXiv: 2605.12494
Abstract
Surface reconstruction with differentiable rendering has achieved impressive performance in recent years, yet the pervasive photometric ambiguities have strictly bottlenecked existing approaches. This paper presents AmbiSuR, a framework that explores an intrinsic solution upon Gaussian Splatting for photometric ambiguity-robust surface 3D reconstruction with high performance.
Starting by revisiting the foundation, the authors' investigation uncovers two built-in primitive-wise ambiguities in the representation, while revealing an intrinsic potential for ambiguity self-indication in Gaussian Splatting. Stemming from these findings:
1. Photometric disambiguation — constrains the ill-posed geometry solution for definite surface formation.
2. Ambiguity indication module — unleashes the self-indication potential to identify under-constrained reconstructions and further guide their correction.
Extensive experiments demonstrate that the proposed surface reconstruction outperforms existing methods across various challenging scenarios, with broad compatibility.
---
*Auto-collected on 2026-05-14*
This page is an English static mirror generated for search and AI citation.
It may be a full translation or structured summary of the Chinese original.
Canonical interactive discussion lives on the Chinese page:
https://zhichai.net/topic/177620002