VAE-CoGAN: Unpaired image-to-image translation for low-level vision

作     者：Zhang, Juan Lang, Xiaoqi Huang, Bo Jiang, Xiaoyan 

作者机构：Shanghai Univ Engn Sci Shanghai 201620 Peoples R China 

出 版 物：《SIGNAL IMAGE AND VIDEO PROCESSING》 (信号，图像与视频处理)

年 卷 期：2023年第17卷第4期

页      面：1019-1026页

核心收录：

学科分类：0808[工学-电气工程] 1002[医学-临床医学] 0809[工学-电子科学与技术（可授工学、理学学位）] 08[工学] 

基　　金：National Natural Science Foundation of China [61702322  61772328  61801288] 

主　　题：Dehaze Derain Generative adversarial networks Variational autoencoder 

摘      要：Low-level vision problems, such as single image haze removal and single image rain removal, usually restore a clear image from an input image using a paired dataset. However, for many problems, the paired training dataset will not be available. In this paper, we propose an unpaired image-to-image translation method based on coupled generative adversarial networks (CoGAN) called VAE-CoGAN to solve this problem. Different from the basic CoGAN, we propose a shared-latent space and variational autoencoder (VAE) in framework. We use synthetic datasets and the real-world images to evaluate our method. The extensive evaluation and comparison results show that the proposed method can be effectively applied to numerous low-level vision tasks with favorable performance against the state-of-the-art methods.