DOI number:  10.1111/cgf.14165

Affiliation of Author(s):  BNRist, Department of Computer Science and Technology, Tsinghua University China

Journal:  Computer Graphics Forum

Funded by:  NSFC (61521002); Beijing Higher Institution ERC; Tsinghua-Tencent Joint Lab

Abstract:  Abstract Image‐to‐image translation has been widely studied. Since real‐world images can often be described by multiple attributes, it is useful to manipulate them at the same time. However, most methods focus on transforming between two domains, and when they chain multiple single attribute transform networks together, the results are affected by the order of chaining, and the performance drops with the out‐of‐domain issue for intermediate results. Existing multi‐domain transfer methods mostly manipulate multiple attributes by adding a list of attribute labels to the network feature, but they also suffer from interference of different attributes, and perform worse when multiple attributes are manipulated. We propose a novel approach to multi‐attribute image‐to‐image translation using several parallel latent transform networks, where multiple attributes are manipulated in parallel and simultaneously, which eliminates both issues. To avoid the interference of different attributes, we introduce a novel soft independence constraint for the changes caused by different attributes. Extensive experiments show that our method outperforms state‐of‐the‐art methods.

Note:  Funding: National Natural Science Foundation of China (61521002); Research Grant of Beijing Higher Institution Engineering Research Center; Tsinghua-Tencent Joint Laboratory for Internet Innovation Technology; Other ISSNs: 1467-8659

Co-author:  Yu-Kun Lai

First Author:  Sen-Zhe Xu

Indexed by:  Journal Article

Discipline:  Engineering

Document Type:  J

Volume:  39

Issue:  7

Page Number:  531–542

ISSN No.:  0167-7055

Translation or Not:  no

Links to published journals:  https://onlinelibrary.wiley.com/doi/10.1111/cgf.14165