A Dual-Tree Complex Wavelet Transform-Based Model for Low-Illumination Image Enhancement  

作　　者：GUAN Yurong Muhammad Aamir Ziaur Rahman Zaheer Ahmed Dayo Waheed Ahmed Abro Muhammad Ishfaq HU Zhihua 

机构地区：[1]Department of Computer Science,Huanggang Normal University,Huanggang 438000,Hubei,China

出　　处：《Wuhan University Journal of Natural Sciences》2021年第5期405-414,共10页武汉大学学报（自然科学英文版）

基　　金：Supported by Teaching Team Project of Hubei Provincial Department of Education(203201929203);the Natural Science Foundation of Hubei Province(2021CFB316);New Generation Information Technology Innovation Project Ministry of Education(20202020ITA05022);Hundreds of Schools Unite with Hundreds of Counties-University Serving Rural Revitalization Science and Technology Support Action Plan(BXLBX0847)。

摘　　要：Image enhancement is a monumental task in the field of computer vision and image processing.Existing methods are insufficient for preserving naturalness and minimizing noise in images.This article discusses a technique that is based on wavelets for optimizing images taken in low-light.First,the V channel is created by mapping an image’s RGB channel to the HSV color space.Second,the acquired V channel is decomposed using the dual-tree complex wavelet transform(DT-CWT)in order to recover the concentrated information within its high and low-frequency subbands.Thirdly,an adaptive illumination boost technique is used to enhance the visibility of a low-frequency component.Simultaneously,anisotropic diffusion is used to mitigate the high-frequency component’s noise impact.To improve the results,the image is reconstructed using an inverse DT-CWT and then converted to RGB space using the newly calculated V.Additionally,images are white-balanced to remove color casts.Experiments demonstrate that the proposed approach significantly improves outcomes and outperforms previously reported methods in general.

关 键 词：image enhancement dual-tree complex wavelet transform(DT-CWT) anisotropic diffusion low-light images 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]