SplatBright: Generalizable Low-Light Scene Reconstruction from Sparse Views via Physically-Guided Gaussian Enhancement

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Wen, Yue, Song, Liang, Wang, Hesheng
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917160527331328
author Wen, Yue
Song, Liang
Wang, Hesheng
author_facet Wen, Yue
Song, Liang
Wang, Hesheng
contents Low-light 3D reconstruction from sparse views remains challenging due to exposure imbalance and degraded color fidelity. While existing methods struggle with view inconsistency and require per-scene training, we propose SplatBright, which is, to our knowledge, the first generalizable 3D Gaussian framework for joint low-light enhancement and reconstruction from sparse sRGB inputs. Our key idea is to integrate physically guided illumination modeling with geometry-appearance decoupling for consistent low-light reconstruction. Specifically, we adopt a dual-branch predictor that provides stable geometric initialization of 3D Gaussian parameters. On the appearance side, illumination consistency leverages frequency priors to enable controllable and cross-view coherent lighting, while an appearance refinement module further separates illumination, material, and view-dependent cues to recover fine texture. To tackle the lack of large-scale geometrically consistent paired data, we synthesize dark views via a physics-based camera model for training. Extensive experiments on public and self-collected datasets demonstrate that SplatBright achieves superior novel view synthesis, cross-view consistency, and better generalization to unseen low-light scenes compared with both 2D and 3D methods.
format Preprint
id arxiv_https___arxiv_org_abs_2512_18655
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SplatBright: Generalizable Low-Light Scene Reconstruction from Sparse Views via Physically-Guided Gaussian Enhancement
Wen, Yue
Song, Liang
Wang, Hesheng
Computer Vision and Pattern Recognition
Low-light 3D reconstruction from sparse views remains challenging due to exposure imbalance and degraded color fidelity. While existing methods struggle with view inconsistency and require per-scene training, we propose SplatBright, which is, to our knowledge, the first generalizable 3D Gaussian framework for joint low-light enhancement and reconstruction from sparse sRGB inputs. Our key idea is to integrate physically guided illumination modeling with geometry-appearance decoupling for consistent low-light reconstruction. Specifically, we adopt a dual-branch predictor that provides stable geometric initialization of 3D Gaussian parameters. On the appearance side, illumination consistency leverages frequency priors to enable controllable and cross-view coherent lighting, while an appearance refinement module further separates illumination, material, and view-dependent cues to recover fine texture. To tackle the lack of large-scale geometrically consistent paired data, we synthesize dark views via a physics-based camera model for training. Extensive experiments on public and self-collected datasets demonstrate that SplatBright achieves superior novel view synthesis, cross-view consistency, and better generalization to unseen low-light scenes compared with both 2D and 3D methods.
title SplatBright: Generalizable Low-Light Scene Reconstruction from Sparse Views via Physically-Guided Gaussian Enhancement
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.18655