Attention-Guided Diffusion Model for Adaptive Low-Light Image Enhancement

Low-light image enhancement is a fundamental task in computer vision, playing a critical role in applications such as autonomous driving, surveillance, and aerial imaging. However, low-light images often suffer from severe noise, loss of detail, and poor contrast, which degrade visual quality and hinder downstream tasks. To address this limitation, we propose an adaptive noise modulation framework that employs an attention-guided noise density mechanism. Unlike previous approaches, our method utilizes attention maps to analyze the importance of different regions within an image, dynamically adjusting noise density accordingly. Critical details are preserved in high-texture areas by applying sparser noise, while denser noise is used in less significant regions to effectively enhance darker areas. This structure-aware noise modulation allows our model to progressively refine image features, achieving stable and high-fidelity restoration. Extensive experiments on multiple low-light datasets demonstrate that our approach surpasses state-of-the-art methods in both quantitative evaluations and perceptual quality, offering a robust and efficient solution for real-world low-light image enhancement.