基于U-net++网络的弱光图像增强方法

doi:10.11896/jsjkx.210300111

计算机科学 ›› 2021, Vol. 48 ›› Issue (11A): 278-282.doi: 10.11896/jsjkx.210300111

• 图像处理& 多媒体技术 • 上一篇下一篇

基于U-net++网络的弱光图像增强方法

李华基, 程江华, 刘通, 程榜, 赵康成

国防科技大学电子科学学院长沙410073

出版日期:2021-11-10 发布日期:2021-11-12
通讯作者: 程江华(jianghua_cheng@nudt.edu.cn)
作者简介:1046866485@qq.com
基金资助:
湖南省自然科学基金(2020JJ4670)

Low-light Image Enhancement Method Based on U-net++ Network

LI Hua-ji, CHENG Jiang-hua, LIU Tong, CHENG Bang, ZHAO Kang-cheng

College of Electronic Science,National University of Defense Technology,Changsha 410073,China

Online:2021-11-10 Published:2021-11-12
About author:LI Hua-ji,born in 1996,postgraduate.His main research interests include computer vision and intelligent information processing.
CHENG Jiang-hua,born in 1979,Ph.D,professor,master supervisor.His main research interests include computer vision and intelligent information processing.
Supported by:
Natural Science Foundation of Hunan Province(2020JJ4670).

摘要/Abstract

摘要： 弱光图像增强是计算机视觉中最具挑战性的任务之一,现有算法存在亮度不均、对比度低、颜色失真和噪声严重等问题。文中提出了一种基于改进U-net++网络实现更为自然的暗光增强网络框架。首先,输入弱光图像至改进U-net++网络,利用各层密集连接以增强不同层次图像特征的关联性;其次,把各层次图像特征融合后输入卷积网络层进行细节重建。实验结果证明,该方法在提高图像亮度的同时,更好地恢复了弱光图像的细节特征,并且生成正常光图像的颜色特征更接近自然。在PASCAL VOC测试集上的测试结果显示结构相似度(SSIM)和峰值信噪比(PSNR)两个重要指标分别为0.87和26.36,比同类最优算法分别高出18.6%和11.4%。

关键词: U-net++网络, 密集连接, 弱光增强, 细节重建

Abstract: Low-light image enhancement is one of the most challenging tasks in computer vision.The current algorithms have some problems,such as uneven brightness,low contrast,color distortion and serious noise.In this paper,a more natural dark light enhanced network framework based on improved U-net++ network is proposed.First of all,the low light image is input to the improved U-net++ network,and the dense connection of each layer is used to enhance the correlation of different levels of image features.Secondly,the image features of each level are fused and input to the convolution network layer for detail reconstruction.The experimental results show that the proposed method not only improves the brightness of the image,but also restores the detail features of the low light image better,and the color feature of the normal light image is closer to the nature.Tests on the PASCAL VOC test set show that the two important indicators,structural similarity (SSIM) and peak signal-to-noise ratio (PSNR),are 0.87 and 26.36,which are 18.6% and 11.4% higher than similar optimal algorithms respectively.

Key words: Dense connection, Detail reconstruction, Low-light enhancement, U-net++ network

中图分类号:

TP391

李华基, 程江华, 刘通, 程榜, 赵康成. 基于U-net++网络的弱光图像增强方法[J]. 计算机科学, 2021, 48(11A): 278-282. https://doi.org/10.11896/jsjkx.210300111

LI Hua-ji, CHENG Jiang-hua, LIU Tong, CHENG Bang, ZHAO Kang-cheng. Low-light Image Enhancement Method Based on U-net++ Network[J]. Computer Science, 2021, 48(11A): 278-282. https://doi.org/10.11896/jsjkx.210300111

参考文献

[1]ARICI T,DIKBAS S,ALTUNBASAK Y.A histogram modification frame-work and its application for image contrast enhancement [J].IEEE Transactions on Image Processing,2009,18(9):1921-1935.
[2]NAKAI K,HOSHI Y,TAGUCHI A.Color image contrast enhacement method based on differential intensity/saturation gray-levels histograms[C]//IEEE International Symposiumon Intelligent Signal Processing and Communications Systems (ISPACS).2013:445-449.
[3]WANG S H,ZHENG J,HU H M,et al.Naturalness preserved enhancement algorithm for nonuniform illumination images [J].IEEE Transactions on Image Processing,2013,22(9):3538-3548.
[4]FU X Y,ZENG D L,HUANG Y,et al.A weighted variational model for simultaneous reflectance and illumination estimation[C]//IEEE Conference on Computer Vision and Pattern Recognition.2016:2782-2790.
[5]LORE K G,AKINTAYO A,SARKAR S.Llnet:A deep autoencoder approach to natural low-light image enhancement [J].Pattern Recognition,2017,61:650-662.
[6]TAO L,ZHU C,XIANG G Q,et al.Llcnn:A convolutional neural network for low-light image enhancement[C]//IEEE Visual Communications and Image Processing (VCIP).2017:1-4.
[7]SHEN L,YUE Z,FENG F,et al.MSR-net:Low-light ImageEnhancement Using Deep Convolutional Network[EB/OL].[2017-11-07].https://arxiv.org/abs/1711.02488.
[8]WEI C,WANG W,YANG W,et al.Deep Retinex Decomposition for Low-Light Enhancement[C]//British Machine Vision Conference.2018.
[9]DABOV K,FOI A,KATKOVNIK V,et al.Image Denoising by Sparse 3D Transform-Domain Collaborative Filtering [J].IEEE Transactions on Image Processing,2007,16(8):2080-2095.
[10]WANG W,CHEN W,YANG W,et al.GLADNet:Low-LightEnhancement Network with Global Awareness[C]//IEEE International Conference on Automatic Face & Gesture Recognition.2018:751-755.
[11]CHEN C,CHEN Q,XU J,et al.Learning to See in the Dark[C]//IEEE Conference on Computer Vision and Pattern Recognition.2018:3291-3300.
[12]ZHOU Z,SIDDIQUEE M M R,TAJBAKHSH N,et al.UNet++:a nested U-Net architecture for medical image segmentation[EB/OL].[2019-10-16].https://arxiv.org/pdf/1807.10165v1.pdf.
[13]WANG Z,BOVIK A C,SHEIKH H R,et al.Image quality assessment:from error visibility to structural similarity [J].IEEE Transactions on Image Processing,2004,13(4):600-612.
[14]HE K,ZHANG X,REN S,et al.Deep Residual Learning for Image Recognition[C]//IEEE Conference on Computer Vision and Pattern Recognition (CVPR).2016:770-778.
[15]LV F,LU F,WU J,et al.MBLLEN:Low-Light Image/Video Enhancement Using CNNs[C]//BMVC.2018:220.
[16]HAMID R S,ALAN C B.Image information and visual quality [J].IEEE Transactions on Image Processing,2006,15(2):430-444.
[17]WANG S,ZHENG J,HU H,et al.Naturalness preserved en-hancement algorithm for non-uniform illumination images [J].IEEE Transactions on Image Processing,2013,22(9):3538-3548.
[18]MITTAL A,SOUNDARARAJAN R,BOVIK A.Making acompletely blind image quality analyzer[J].IEEE Signal Processing Letters,2013,20(3):209-212.
[19]YING Z Q,LI G,REN Y R,et al.A new image contrast enhancement algorithm using exposure fusion framework[C]//International Conference on Computer Analysis of Images and Patterns.Springer,2017:36-46.
[20]GUO X J,LI Y,LING H B.Lime:Low-light image enhancement via illumination map estimation [J].IEEE Transactions on Image Processing,2017,26(2):982-993.
[21]FU X Y,ZENG D L,HUANG Y,et al.A fusion-based enhancing method for weakly illuminated images[J].Signal Processing,2016,129:82-96.
[22]ZHANG Y H,ZHANG J W,GUO X J.Kindling the Darkness:A Practical Low-light Image Enhancer[C]//27th ACM International Conference on Multimedia.2019:1632-1640.
[23]GUO C,LI C,GUO J,et al.Zero-Reference Deep Curve Estimation for Low-Light Image Enhancement[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).IEEE,2020:1777-1786.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

基于U-net++网络的弱光图像增强方法

Low-light Image Enhancement Method Based on U-net++ Network

PDF (PC)

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 3

Metrics

本文评价

推荐阅读 0

[1]	王丽芳, 王蕊芳, 蔺素珍, 秦品乐, 高媛, 张晋. 基于双残差超密集网络的多模态医学图像融合 Multimodal Medical Image Fusion Based on Dual Residual Hyper Densely Networks 计算机科学, 2021, 48(2): 160-166. https://doi.org/10.11896/jsjkx.200400095
[2]	钱小梅,刘嘉勇,程芃森. 基于密集连接卷积神经网络的远程监督关系抽取 Distant Supervised Relation Extraction Based on Densely Connected Convolutional Networks 计算机科学, 2020, 47(2): 157-162. https://doi.org/10.11896/jsjkx.190100167
[3]	郑诚, 薛满意, 洪彤彤, 宋飞豹. 用于短文本分类的DC-BiGRU_CNN模型 DC-BiGRU_CNN Model for Short-text Classification 计算机科学, 2019, 46(11): 186-192. https://doi.org/10.11896/jsjkx.180901702