TY - JOUR
T1 - Single-image de-raining with feature-supervised generative adversarial network
AU - Xiang, Peng
AU - Wang, Lei
AU - Wu, Fuxiang
AU - Cheng, Jun
AU - Zhou, Mengchu
N1 - Funding Information:
Manuscript received January 23, 2019; revised March 2, 2019; accepted March 3, 2019. Date of publication March 8, 2019; date of current version March 20, 2019. This work was supported in part by the National Key R&D Program of China (2018YFB1308000), in part by the National Natural Science Foundation of China (61772508 and U1713213); in part by Shenzhen Technology Project (JCYJ20170413152535587, JCYJ 20170307164023599, and JSGG20170823091924128); in part by CAS Key Technology Talent Program, Guangdong Technology Program (2016B010108010, 2016B010125003, and 2017B010110007); and in part by Shenzhen Engineering Laboratory for 3D Content Generating Technologies ([2017] 476). The associate editor coordinating the review of this manuscript and approving it for publication was Dr. Kai Liu. (Corresponding author: Lei Wang.) P. Xiang, L. Wang, F. Wu, and J. Cheng are with the Guangdong Provincial Key Laboratory of Robotics and Intelligent System, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen 518055, China, and also with The Chinese University of Hong Kong, Hong Kong (e-mail:,xiangpeng@126.com; lei.wang1@siat.ac.cn; fx.wu1@siat.ac.cn; jun.cheng@siat.ac.cn).
PY - 2019/5
Y1 - 2019/5
N2 - De-raining, which aims at rain-steak removal from images, is a practical task in computer vision. However, it is difficult due to its ill-posed nature. In this letter, we propose a deep neural network architecture, feature-supervised generative adversarial network (FS-GAN) for single-image rain removal. Its main idea is to train a generative adversarial network (GAN) for which the supervision from ground truth is imposed on different layers of the generator network. We design a feature-supervised generator, a discriminator, an optimization target, as well as the detailed structure of FS-GAN. Experiments show that the proposed FS-GAN achieves better performance than state-of-the-art de-raining methods on both synthetic and real-world images in terms of quantitative and visual quality.
AB - De-raining, which aims at rain-steak removal from images, is a practical task in computer vision. However, it is difficult due to its ill-posed nature. In this letter, we propose a deep neural network architecture, feature-supervised generative adversarial network (FS-GAN) for single-image rain removal. Its main idea is to train a generative adversarial network (GAN) for which the supervision from ground truth is imposed on different layers of the generator network. We design a feature-supervised generator, a discriminator, an optimization target, as well as the detailed structure of FS-GAN. Experiments show that the proposed FS-GAN achieves better performance than state-of-the-art de-raining methods on both synthetic and real-world images in terms of quantitative and visual quality.
KW - Rain removal
KW - convolutional neural networks
KW - generative adversarial network
UR - http://www.scopus.com/inward/record.url?scp=85063594105&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85063594105&partnerID=8YFLogxK
U2 - 10.1109/LSP.2019.2903874
DO - 10.1109/LSP.2019.2903874
M3 - Article
AN - SCOPUS:85063594105
SN - 1070-9908
VL - 26
SP - 650
EP - 654
JO - IEEE Signal Processing Letters
JF - IEEE Signal Processing Letters
IS - 5
M1 - 8663315
ER -