Low-light Image Enhancement via Multi-scale Attention combined with Fourier Transform

2026-07-27

AI Summary

Existing low-light image enhancement methods struggle to accurately model real-world illumination and recover fine textures. To address this limitation, this work proposes a single-stage U-shaped network that innovatively integrates frequency-domain magnitude priors with a multi-scale attention mechanism. The architecture further incorporates multi-shape collaborative attention and a lightweight design, enabling efficient embedding of high-dimensional texture features under frequency-domain supervision. Extensive experiments demonstrate that the proposed method significantly outperforms current state-of-the-art approaches across multiple benchmarks, including LOL, SID, SMID, and SDSD. Notably, on the SDSD-outdoor dataset, it achieves a PSNR of 41.76 dB—an improvement of 11.92 dB—and an SSIM of 0.988, representing a 13.80% gain over prior methods.

Abstract

Low-light image enhancement (LLIE) aims to improve image quality and clarity in diverse and demanding low-illumination environments. However, existing deep learning-based LLIE methods struggle to accurately capture real-world illumination and restore texture details, largely because their algorithmic strengths remain underutilized. To address these issues, we present a supervised frequency domain deep learning network for LLIE, named multi-scale attention combined with the Fourier transform (MSFT) which adopts a U-shaped, one-stage architecture that infuses guidance from low-light images into the network by channeling it through multi-scale attention. We further fuse the amplitude information from priori channels with that of the low-light image in MSFT's self-created module, and carry out multi-scale guidance along with the network. Subsequently, to better enhance the faint feature, such as fine content and textures, and to better fuse global context confidence in the decoding stage, we separately introduce a multi-shape synergistic attention and a lightweight network that effectively integrate information in high-dimensional space to embed into the superlative feature space channel containing rich texture information. Extensive experiments conducted on LOL, SID, SMID, and SDSD datasets demonstrate that MSFT significantly outperforms state-of-the-art competitors. For example, compared with Retinexformer, our method achieves a peak signal-to-noise ratio of up to 41.76 decibels on the SDSD-outdoor dataset with an increase of 11.92 decibels and a structural similarity index of 0.988 with a 13.80% improvement.

Problem

• Low-light Image Enhancement
• Texture Detail Restoration
• Illumination Estimation
• Deep Learning
• Image Quality

Innovation

• multi-scale attention
• Fourier transform
• low-light image enhancement
• frequency domain learning
• synergistic attention