一种改进的生成对抗网络的图像去模糊方法  

An Improved Image Deblurring Method of Generative Adversarial Network

作　　者：侯坤 李佳[1] 沈宇慧 HOU Kun;LI Jia;SHEN Yuhui(School of Electronics and Information Engineering,Nanjing University of Information Science and Technology,Nanjing Jiangsu 210044,China)

机构地区：[1]南京信息工程大学电子与信息工程学院,江苏南京210044

出　　处：《信息与电脑》2021年第19期58-60,共3页Information & Computer

摘　　要：针对利用现有图像去模糊方法复原图像存在振铃效应、细节模糊等问题,笔者提出一种改进的生成对抗网络的图像去模糊方法。该方法利用改进的特征金字塔网络生成复原图像,并通过判别器进行对抗训练从而获得去模糊图像。实验结果表明,该方法能减轻复原图像中的振铃效应,达到恢复图像细节的目的。Aiming at the problems such as ringing effect and detail blurring in the existing image deblurring methods, an improved image deblurring method based on generating adversative network was proposed. The improved feature pyramid network is used to generate the restored image, and the deblurring image is obtained through the antagonistic training of the discriminator. Experimental results show that this method can reduce the ringing effect in the restored image and achieve the purpose of restoring image details.

关 键 词：图像去模糊 生成对抗网络 特征金字塔 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP183[自动化与计算机技术—计算机科学与技术]