基于无配对生成对抗网络的图像超分辨率重建  

Image Super-resolution Based on No Match Generative Adversarial Network

作　　者：李学相[1] 曹淇 刘成明[1] LI Xuexiang;CAO Qi;LIU Chengming(School of Software,Zhengzhou University,Zhengzhou 450002,China)

机构地区：[1]郑州大学软件学院,河南郑州450002

出　　处：《郑州大学学报（工学版）》2021年第5期1-6,共6页Journal of Zhengzhou University（Engineering Science）

基　　金：国家重点研发计划项目(2020YFB1712401)。

摘　　要：针对基于生成对抗网络的图像超分辨率重建方法依赖配对数据集训练且结果不稳定的问题,提出了一个新的基于无配对图像的模型NM-SRGAN。首先,通过使用循环生成对抗网络作预处理模块,使模型可以不依赖配对数据集进行训练且获得更好的输入图像,同时该模型取消了BN层的使用,解决了结果不稳定的问题。然后,使用了协方差矩阵捕捉图像的二阶信息,增加了二阶损失函数,更加注重于捕捉图像细节区域部分的变化。最后,通过使用新的VGG损失函数提升了图像的边缘纹理细节。对提出的NM-SRGAN模型在4个标准数据集上进行测试评估,使用客观评价标准对结果图进行评价,NM-SRGAN模型较目前若干先进模型中的最佳峰值信噪比分别提升了0.19、0.03、0.13、0.02 d B,在4个数据集上的评价值均达到最高。实验结果表明,该模型在稳定性、图像质量和细节方面较经典算法均有较好的提升。Image super-resolution reconstruction based on generative adversarial networks( GAN) is subject to the dataset training with an unstable result. To solve this problem,a new NM-SRGAN model is established.The cycle-gan is firstly used as the preprocess module to make the model free from the dataset for training with better input of the image,and the model cancels BN layer to solve the unstable results. Besides,covariance matrix is adopted to capture the second-order information of the image,and second-order loss function is added with a focus on the changes of the image details. The new VGG loss function is used to improve the marginal texture of the image. The proposed NM-SRGAN model is verified by four standard datasets,and the resulting images are assessed by the objective evaluation indices. Compared with the existing models,NM-SRGAN model has an improved evaluation value of 0. 19,0. 03,0. 13,and 0. 02 d B,respectively,reaching up to the maximum among the four datasets. Results show that the proposed method,compared with traditional algorithms,has achieved better improvements in stability and image quality with better details.

关 键 词：超分辨率 深度学习 生成对抗网络 无配对 二阶统计量 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]