中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
热门
Asymmetric GAN for Unpaired Image-to-Image Translation

文献类型:期刊论文

作者Li, Yu1,4; Tang, Sheng1; Zhang, Rui1,4; Zhang, Yongdong1; Li, Jintao1; Yan, Shuicheng2,3
刊名IEEE TRANSACTIONS ON IMAGE PROCESSING
出版日期2019-12-01
卷号28期号:12页码:5881-5896
ISSN号1057-7149
关键词Generative adversarial networks cycle GAN asymmetric GAN image-to-image translation unpaired translation
DOI10.1109/TIP.2019.2922854
英文摘要Unpaired image-to-image translation problem aims to model the mapping from one domain to another with unpaired training data. Current works like the well-acknowledged Cycle GAN provide a general solution for any two domains through modeling injective mappings with a symmetric structure. While in situations where two domains are asymmetric in complexity, i.e., the amount of information between two domains is different, these approaches pose problems of poor generation quality, mapping ambiguity, and model sensitivity. To address these issues, we propose Asymmetric GAN (AsymGAN) to adapt the asymmetric domains by introducing an auxiliary variable (aux) to learn the extra information for transferring from the information-poor domain to the information-rich domain, which improves the performance of state-of-the-art approaches in the following ways. First, aux better balances the information between two domains which benefits the quality of generation. Second, the imbalance of information commonly leads to mapping ambiguity, where we are able to model one-to-many mappings by tuning aux, and furthermore, our aux is controllable. Third, the training of Cycle GAN can easily make the generator pair sensitive to small disturbances and variations while our model decouples the ill-conditioned relevance of generators by injecting aux during training. We verify the effectiveness of our proposed method both qualitatively and quantitatively on asymmetric situation, label-photo task, on Cityscapes and Helen datasets, and show many applications of asymmetric image translations. In conclusion, our AsymGAN provides a better solution for unpaired image-to-image translation in asymmetric domains.
资助项目National Key Research and Development Program of China[2017YFC0820605] ; National Natural Science Foundation of China[61572472] ; National Natural Science Foundation of China[61871004] ; National Natural Science Foundation of China[61525206] ; National Natural Science Foundation of China[2019A010]
WOS研究方向Computer Science ; Engineering
语种英语
出版者IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号WOS:000484306000010
源URL[http://119.78.100.204/handle/2XEOYT63/4753]  
专题中国科学院计算技术研究所期刊论文_英文
通讯作者Tang, Sheng
作者单位1.Chinese Acad Sci, Inst Comp Technol, Key Lab Intelligent Informat Proc, Beijing 100190, Peoples R China
2.Natl Univ Singapore, Dept Elect & Comp Engn, Singapore, Singapore
3.AI Inst, Qihoo 360, Beijing 100025, Peoples R China
4.Univ Chinese Acad Sci, Sch Comp Sci & Technol, Beijing 100049, Peoples R China
推荐引用方式
GB/T 7714
Li, Yu,Tang, Sheng,Zhang, Rui,et al. Asymmetric GAN for Unpaired Image-to-Image Translation[J]. IEEE TRANSACTIONS ON IMAGE PROCESSING,2019,28(12):5881-5896.
APA Li, Yu,Tang, Sheng,Zhang, Rui,Zhang, Yongdong,Li, Jintao,&Yan, Shuicheng.(2019).Asymmetric GAN for Unpaired Image-to-Image Translation.IEEE TRANSACTIONS ON IMAGE PROCESSING,28(12),5881-5896.
MLA Li, Yu,et al."Asymmetric GAN for Unpaired Image-to-Image Translation".IEEE TRANSACTIONS ON IMAGE PROCESSING 28.12(2019):5881-5896.

入库方式: OAI收割

来源:计算技术研究所

浏览0
下载0
收藏0
其他版本

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。