结合注意引导网络的弱光图像增强算法

Low Light Image Enhancement Algorithm with Attention Guided Network

弱光图像增强具有挑战性,不仅需要考虑亮度恢复,还需要考虑色彩失真和噪声等复杂问题。简单地调整弱光图像的亮度将不可避免的放大这些伪影。为了解决这些难题,一种带有注意引导分支的端到端弱光增强网络(attention guided low light enhancement network,AGNet)被提出。AGNet由注意引导网络和弱光增强网络两部分组成。注意引导网络被用来学习弱光图像中的照度-注意映射,并将其应用于弱光增强网络,以指导图像亮度增强和去噪任务。弱光增强网络由多尺度卷积和残差块构成,通过特征金字塔结构从多个尺度来提取弱光图像中的细节和纹理特征。此外,网络中还引入了多尺度色彩矫正模块(multi-scale color recalibration module,MCRM),以进一步增强了输出图像的颜色和对比度。实验结果表明,AGNet在主流弱光数据集上(LOL-v1和LOL-v2-synthetic)不仅在客观指标上领先(两个数据集的PSNR提高了2.13/2.52),而且在主观比较上也具有优势。

Low-light image enhancement is a challenging task that requires consideration of various complex issues,including brightness restoration,color distortion,and noise reduction.Simple adjustments to the brightness of low-light images inevitably magnify these artifacts.To address these challenges,an attention guided low light enhancement network(AGNet)was proposed.AGNet consists of two parts:an attention guided network and a low-light enhancement network.The attention guided network is used to learn the illuminance-attention mapping in low-light images and applied in low-light enhancement network to guide the brightness enhancement and denoising tasks.The low-light enhancement network is composed of multi-scale convolution and residual blocks,which extract detail and texture features from low-light images through feature pyramid structure.Furthermore,a multi-scale color recalibration module(MCRM)is introduced in the network to further enhance the color and contrast of the output image.Experimental results demonstrate that AGNet not only outperforms other methods on major low-light datasets(LOL-v1 and LOL-v2-synthetic)in terms of objective metrics(the PSNR of both datasets improved by 2.13/2.52),but it also has subjective advantages in comparison.