LLGS: Unsupervised Gaussian Splatting for Image Enhancement and Reconstruction in Pure Dark Environment

2025-03-24

AI Summary

To address critical challenges in 3D Gaussian Splatting (3DGS) under pure darkness—including deficient color representation, multi-view inconsistency, and poor generalization—this paper proposes the first end-to-end, unsupervised 3DGS optimization framework for low-light modeling. Our method introduces: (1) M-Color decomposable Gaussian representation, enabling differentiable, disentangled modeling of geometry, illumination, and chromaticity; and (2) a zero-knowledge-guided enhancement mechanism that jointly optimizes geometry and illumination in an unsupervised manner, incorporating direction-aware enhancement without paired data or pretrained priors. Evaluated on real-world low-light datasets, our approach achieves state-of-the-art performance in both image enhancement (PSNR/SSIM) and 3D reconstruction (Chamfer Distance), marking the first solution to deliver multi-view consistent, high-fidelity, and fully unsupervised 3D modeling under extreme low-light conditions.

Technology Category

• Computer Vision: 3D Computer Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

3D Gaussian Splatting has shown remarkable capabilities in novel view rendering tasks and exhibits significant potential for multi-view optimization.However, the original 3D Gaussian Splatting lacks color representation for inputs in low-light environments. Simply using enhanced images as inputs would lead to issues with multi-view consistency, and current single-view enhancement systems rely on pre-trained data, lacking scene generalization. These problems limit the application of 3D Gaussian Splatting in low-light conditions in the field of robotics, including high-fidelity modeling and feature matching. To address these challenges, we propose an unsupervised multi-view stereoscopic system based on Gaussian Splatting, called Low-Light Gaussian Splatting (LLGS). This system aims to enhance images in low-light environments while reconstructing the scene. Our method introduces a decomposable Gaussian representation called M-Color, which separately characterizes color information for targeted enhancement. Furthermore, we propose an unsupervised optimization method with zero-knowledge priors, using direction-based enhancement to ensure multi-view consistency. Experiments conducted on real-world datasets demonstrate that our system outperforms state-of-the-art methods in both low-light enhancement and 3D Gaussian Splatting.

Problem

• Enhancing images in pure dark environments without pre-trained data
• Ensuring multi-view consistency in low-light 3D reconstruction
• Overcoming color representation limitations in low-light Gaussian Splatting

Innovation

• Unsupervised multi-view stereoscopic system
• Decomposable Gaussian representation M-Color
• Direction-based enhancement for consistency