Residual learning based deep generative networks have achieved promising performance in im-age enhancement.However,due to the large color gap between a low-quality image and its high-quality version,the identical mapping in conventional residual learning cannot explore the elaborate detail differences,resulting in color deviations and texture losses in enhanced images.To address this issue,an innovative non-identical residual learning architecture is proposed,which views image enhancement as two complementary branches,namely a holistic color adjustment branch and a fine-grained residual generation branch.In the holistic color adjustment,an adjusting map is calculated for each input low-quality image,in order to regulate the low-quality image to the high-quality repre-sentation in an overall way.In the fine-grained residual generation branch,a novel attention-aware recursive network is designed to generate residual images.This design can alleviate the overfitting problem by reusing parameters and promoting the network's adaptability for different input condi-tions.In addition,a novel dynamic multi-level perceptual loss based on the error feedback ideology is proposed.Consequently,the proposed network can be dynamically optimized by the hybrid per-ceptual loss provided by a well-trained VGG,so as to improve the perceptual quality of enhanced images in a guided way.Extensive experiments conducted on publicly available datasets demonstrate the state-of-the-art performance of the proposed method.

HU Ruiguang;HUANG Li

Beijing Aerospace Automatic Control Institute,Beijing 100854,P.R.China

高技术通讯（英文版）

[1]HU Ruiguang;HUANG Li-.Non-identical residual learning for image enhancement via dynamic multi-level perceptual loss)[J].高技术通讯（英文版）,2022(02):142-152

Non,identical,residual,learning,enhancement,multi,level,perceptual,Residual,deep,generative,networks,have,achieved,promising,performance,However,due,large,color,gap,between,low,quality,its,high,version,mapping,conventional,cannot,explore,elaborate,detail,differences,resulting,deviations,texture,losses,enhanced,images,To,address,this,issue,innovative,architecture,proposed,which,views,complementary,branches,namely,holistic,adjustment,fine,grained,generation,In,adjusting,calculated,each,input,order,regulate,repre,sentation,overall,way,novel,attention,aware,recursive,designed,generate,This,alleviate,overfitting,problem,by,reusing,parameters,promoting,adaptability,different,condi,addition,error,feedback,ideology,Consequently,dynamically,optimized,hybrid,provided,well,trained,VGG,so,improve,guided,Extensive,experiments,conducted,publicly,available,datasets,demonstrate,state,art,method

0.537435