Scene-Adaptive Nonlinear Tone Curves for Pseudo Ground-Truth Generation in Low-Light 3D Gaussian Splatting

2026-06-10

AI Summary

This work addresses the challenge of novel view synthesis under low-light, multi-view conditions, where high noise levels, weak details, and limited dynamic range hinder reconstruction quality. Existing pseudo-ground-truth generation methods rely on linear gain adjustments, often causing highlight clipping and insufficient dark-region enhancement. To overcome these limitations, this study introduces nonlinear tone mapping into pseudo-ground-truth generation for 3D reconstruction, proposing a scene-adaptive framework that integrates percentile normalization with adaptive black-level offsetting. Two complementary tone-mapping curves are designed: a bounded exponential (ASE) and a data-driven cubic polynomial (AP3). By replacing only the pseudo-ground-truth module in 3D Gaussian Splatting, the method achieves significant improvements—up to 4.34 dB PSNR gain on LOM and 3.25 dB on RealX3D across 21 scenes—and demonstrates comparable performance between both curves, indicating that the gains stem from the adaptive mechanism rather than the specific curve formulation.

Abstract

Low-light novel view synthesis is challenging because dark multi-view images contain noise, weak structural detail, and compressed dynamic range. Recent 3D Gaussian Splatting (3DGS) methods address these challenges by generating pseudo ground-truth (pseudo-GT) images as supervision targets when paired normal-light references are unavailable. Existing pseudo-GT methods apply a uniform linear gain to all pixels, which clips bright regions while providing insufficient enhancement in dark regions, limiting reconstruction quality. We observe that nonlinear tone mappings, long established in 2D low-light enhancement, have not been explored for pseudo-GT generation in 3D reconstruction. Accordingly, we propose a scene-adaptive nonlinear tone-curve framework that replaces linear pseudo-GT with nonlinear alternatives. The framework introduces percentile-based normalisation for scene-agnostic curve application, a scene-adaptive offset for automatic black-level adjustment, and two complementary curves: Adaptive SoftExp (ASE), a bounded exponential curve, and Adaptive Poly3 (AP3), a data-driven cubic polynomial. The module changes only the pseudo-GT computation and leaves the 3DGS backbone unchanged. Experiments on three benchmarks covering 21 scenes show that both curves consistently outperform the linear baseline with PSNR improvements up to +4.34 dB on LOM and +3.25 dB on RealX3D. Both curves achieve similar performance despite their different mathematical forms, suggesting the improvement is curve-agnostic. Code is available at https://github.com/lvmingzhe/adaptiveToneCurve

Problem

• low-light
• pseudo ground-truth
• 3D Gaussian Splatting
• tone mapping
• novel view synthesis

Innovation

• nonlinear tone mapping
• scene-adaptive
• pseudo ground-truth
• 3D Gaussian Splatting
• low-light enhancement