Illuminating Darkness: Enhancing Real-world Low-light Scenes with Smartphone Images

2025-03-10

AI Summary

To address the performance bottleneck of single-exposure low-light image enhancement caused by the scarcity of high-quality paired data in real-world scenarios, this paper introduces the first large-scale, high-resolution real-world low-light enhancement dataset—comprising 6,425 pairs of 4K+ images captured under illumination levels ranging from 0.1 to 200 lux—and proposes the LC-Tuning Fork Transformer, a luminance-chrominance (LC) decoupled architecture. The model pioneers an LC-decoupled modeling paradigm, incorporating LC cross-attention, LC refinement blocks, and LC-guided supervision to jointly optimize noise suppression and perceptual consistency. Training employs multi-scale non-overlapping patch strategies, while data acquisition and annotation are conducted using dynamic real smartphone sensors. Evaluated on a 400-pair benchmark, our method significantly outperforms state-of-the-art approaches. It demonstrates strong generalization across diverse smartphone hardware and complex indoor/outdoor scenes. Both code and the full dataset are publicly released to advance practical deployment of low-light enhancement.

Technology Category

• Computer Vision: Large Vision Models
• Search and Optimization: Learning to Search
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Security and Privacy: Privacy-enhancing technologies

Abstract

Digital cameras often struggle to produce plausible images in low-light conditions. Improving these single-shot images remains challenging due to a lack of diverse real-world pair data samples. To address this limitation, we propose a large-scale high-resolution (i.e., beyond 4k) pair Single-Shot Low-Light Enhancement (SLLIE) dataset. Our dataset comprises 6,425 unique focus-aligned image pairs captured with smartphone sensors in dynamic settings under challenging lighting conditions (0.1--200 lux), covering various indoor and outdoor scenes with varying noise and intensity. We extracted and refined around 180,000 non-overlapping patches from 6,025 collected scenes for training while reserving 400 pairs for benchmarking. In addition to that, we collected 2,117 low-light scenes from different sources for extensive real-world aesthetic evaluation. To our knowledge, this is the largest real-world dataset available for SLLIE research. We also propose learning luminance-chrominance (LC) attributes separately through a tuning fork-shaped transformer model to enhance real-world low-light images, addressing challenges like denoising and over-enhancement in complex scenes. We also propose an LC cross-attention block for feature fusion, an LC refinement block for enhanced reconstruction, and LC-guided supervision to ensure perceptually coherent enhancements. We demonstrated our method's effectiveness across various hardware and scenarios, proving its practicality in real-world applications. Code and dataset available at https://github.com/sharif-apu/LSD-TFFormer.

Problem

• Enhancing low-light smartphone images in real-world conditions
• Addressing lack of diverse real-world low-light image datasets
• Improving image quality through luminance-chrominance attribute learning

Innovation

• Large-scale high-resolution SLLIE dataset creation
• Luminance-chrominance transformer model for enhancement
• LC cross-attention and refinement blocks integration