光学精密工程2025，Vol.33Issue(13)：2108-2123,16.DOI:10.37188/OPE.20253313.2108
基于光照引导的三阶段低光照图像增强
Three-stage low-light image enhancement based on illumination guidance
摘要
Abstract
Addressing challenges such as severe noise,color deviation,and artifacts in low-light imaging,a novel three-stage low-light image enhancement algorithm based on illumination guidance(IG-TSNet)is proposed.This algorithm synergistically integrates the Fourier domain's capability to capture global image information with the Transformer's strength in modeling long-range dependencies within the spatial do-main,utilizing illumination guidance to ensure coherent enhancement.IG-TSNet comprises three sequen-tial stages.In the pixel-wise enhancement stage,an adaptive parameter adjustment mechanism is intro-duced to improve the global representation of the image.During the Fourier reconstruction stage,illumina-tion priors are employed to optimize both amplitude and phase spectra across two channels following Fouri-er transformation,enabling comprehensive global image reconstruction.In the cross-attention fusion stage,a lightweight dual-path U-shaped network,incorporating a cross-attention fusion module,is de-signed to dynamically align Fourier-reconstructed features with illumination-guidance maps.The proposed IG-TSNet was rigorously evaluated on six benchmark datasets for low-light image enhancement,demon-strating superior performance.Qualitative results confirm that the method effectively enhances underex-posed regions,suppresses noise without introducing artifacts or patchiness,and preserves color fidelity ro-bustly.Quantitative assessments reveal that IG-TSNet achieves state-of-the-art results across nine evalua-tion metrics.On three paired datasets,PSNR values of 26.968 dB,27.880 dB,and 28.939 dB;SSIM values of 0.867,0.882,and 0.947;and LPIPS values of 0.099,0.141,and 0.047 were attained,re-spectively.On three unpaired datasets,BRISQUE scores of 25.67,20.51,and 18.80 and NIQE values of 3.79,4.09,and 4.02 were achieved,respectively.This study offers a viable frequency-spatial joint en-hancement framework,advancing the field of low-light image enhancement.关键词
图像增强/低光照图像/傅里叶频域/交叉注意力/TransformerKey words
image enhancement/low-light image/Fourier frequency domain/cross-attention/transformer分类
信息技术与安全科学引用本文复制引用
张易航,钟寒..基于光照引导的三阶段低光照图像增强[J].光学精密工程,2025,33(13):2108-2123,16.基金项目
中国人民公安大学"双一流"创新研究项目(No.2023SYL07) （No.2023SYL07）
高等学校学科创新引智基地项目(No.B20087) （No.B20087）
