Compression artifacts reduction by improved generative adversarial networks

EURASIP Journal on Image and Video Processing

Qian Sun¹  Haoran Yang¹  Zengshun Zhao²  Heng Qiao³  Dapeng Oliver Wu³  Zhigang Wang⁴ 

[1] 0000 0004 1799 3811, grid.412508.a, College of Electronic and Information Engineering, Shandong University of Science and Technology, 266590, Qingdao, People’s Republic of China;0000 0004 1799 3811, grid.412508.a, College of Electronic and Information Engineering, Shandong University of Science and Technology, 266590, Qingdao, People’s Republic of China;0000 0004 1936 8091, grid.15276.37, Department of Electrical and Computer Engineering, University of Florida, 32611, Gainesville, FL, USA;0000 0004 1761 1174, grid.27255.37, School of Control Science and Engineering, Shandong University, 250061, Jinan, People’s Republic of China;0000 0004 1936 8091, grid.15276.37, Department of Electrical and Computer Engineering, University of Florida, 32611, Gainesville, FL, USA;grid.265025.6, Key Laboratory of Computer Vision and System, Ministry of Education, Tianjin Key Laboratory of Intelligence Computing and Novel Software Technology, Tianjin University of Technology, 300384, Tianjin, People’s Republic of China;

关键词: GANs;    CNN;    Compression artifacts;    JPEG compression;   

DOI  :  10.1186/s13640-019-0465-0

来源: publisher

【 摘 要 】

In this paper, we propose an improved generative adversarial network (GAN) for image compression artifacts reduction task (artifacts reduction by GANs, ARGAN). The lossy compression leads to quite complicated compression artifacts, especially blocking artifacts and ringing effects. To handle this problem, we choose generative adversarial networks as an effective solution to reduce diverse compression artifacts. The structure of “U-NET” style is adopted as the generative network in the GAN. A discriminator network is designed in a convolutional manner to differentiate the restored images from the ground truth distribution. This approach can help improve the performance because the adversarial loss aggressively encourages the output image to be close to the distribution of the ground truth. Our method not only learns an end-to-end mapping from input degraded image to corresponding restored image, but also learns a loss function to train this mapping. Benefit from the improved GANs, we can achieve desired results without hand-engineering the loss functions. The experiments show that our method achieves better performance than the state-of-the-art methods.

【 授权许可】

CC BY