基于深度学习的图像超分辨率重建方法综述  

A summary review of the methods for image super-resolution reconstruction based on deep learning

作　　者：陈文静 唐轶[3] Chen Wen-jing;Tang Yi(Key Laboratory of Spectral Imaging Technology,Xi’an Institute of Optics and Precision Mechanics,Chinese Academy of Sciences,Xi’an 710119,China;University of Chinese Academy of Sciences,Beijing 100049,China;Yunnan Minzu University,Kunming 650500,China)

机构地区：[1]中国科学院西安光学精密机械研究所光谱成像技术重点实验室,陕西西安710119 [2]中国科学院大学,北京100049 [3]云南民族大学数学与计算机科学学院,云南昆明650500

出　　处：《云南民族大学学报（自然科学版）》2019年第6期597-605,共9页Journal of Yunnan Minzu University:Natural Sciences Edition

基　　金：国家自然科学基金(61462096,61866040,61561053)

摘　　要：图像超分辨率重建(super-resolution,SR)是指从观测到的低分辨率图像重建出相应的高分辨率图像,在目标检测、医学成像和卫星遥感等领域都有着重要的应用价值.近年来,随着深度学习的迅速发展,基于深度学习的图像超分辨率重建方法取得了显著的进步.为了把握目前基于深度学习的图像超分辨率重建方法的发展情况和研究热点,对一些最新的基于深度学习的图像超分辨率重建方法进行了梳理,将它们分为两大类(有监督的和无监督的)分别进行阐述.然后,在公开的数据集上,将主流方法的性能进行了对比分析.最后,对基于深度学习的图像超分辨率重建方法进行了总结,并对其未来的研究趋势进行了展望.Image super-resolution reconstruction refers to the reconstruction of corresponding high-resolution images from observed low-resolution images and has important application value in many fields,such as target detection,medical imaging,and satellite remote sensing.In recent years,with the fast development of deep learning,significant progress has been made in the methods for image super-resolution reconstruction based on deep learning.In order to grasp the current development and research hotspots in this field,the latest methods are roughly divided into two categories(supervised and unsupervised)to be described in this review.Then,the performance of the mainstream methods is compared in connection with the publicly available benchmark databases.Finally,the methods for image super-resolution reconstruction based on deep learning are summarized,and future research trends are predicted.

关 键 词：图像超分辨率重建 深度学习 卷积神经网络 生成对抗网络 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]