LLDiffusion: Learning degradation representations in diffusion models for low-light image enhancement

• Published in: Pattern recognition Vol. 166; p. 111628
• Main Authors: Wang, Tao, Zhang, Kaihao, Zhang, Yong, Luo, Wenhan, Stenger, Björn, Lu, Tong, Kim, Tae-Kyun, Liu, Wei
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.10.2025
• Subjects: Degradation aware learning scheme; Degradation representation; Diffusion model; Image enhancement; Degradation aware learning scheme; Image enhancement; Diffusion model; Degradation representation
• ISSN: 0031-3203
• DOI: 10.1016/j.patcog.2025.111628

Current deep learning methods for low-light image enhancement typically rely on pixel-wise mappings using paired data, often overlooking the specific degradation factors inherent to low-light conditions, such as noise amplification, reduced contrast, and color distortion. This oversight can result in suboptimal performance. To address this limitation, we propose a degradation-aware learning framework that explicitly integrates degradation representations into the model design. We introduce LLDiffusion, a novel model composed of three key modules: a Degradation Generation Network (DGNET), a Dynamic Degradation-Aware Diffusion Module (DDDM), and a Latent Map Encoder (E). This approach enables joint learning of degradation representations, with the pre-trained Encoder (E) and DDDM effectively incorporating degradation and image priors into the diffusion process for improved enhancement. Extensive experiments on public benchmarks show that LLDiffusion outperforms state-of-the-art low-light image enhancement methods quantitatively and qualitatively. The source code and pre-trained models will be available at https://github.com/TaoWangzj/LLDiffusion. •We propose degradation-aware diffusion model for low-light image enhancement.•We propose degradation-aware diffusion module for integrating degradation priors.•We develop a real world test dataset for evaluating real-world scenarios.