Multi-scale Generative Adversarial Learning for Facial Attribute Transfer

Yicheng Zhang, Li Song*, Rong Xie, Wenjuan Zhang

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Citations (Scopus)

Abstract

Generative Adversarial Network (GAN) has shown its impressive ability on facial attribute transfer. One crucial part in facial attribute transfer is to retain the identity. To achieve this, most of existing approaches employ the L1 norm to maintain the cycle consistency, which tends to cause blurry results due to the weakness of the L1 loss function. To address this problem, we introduce the Structural Similarity Index (SSIM) in our GAN training objective as the measurement between input images and reconstructed images. Furthermore, we also incorporate a multi-scale feature fusion structure into the generator to facilitate feature learning and encourage long-term correlation. Qualitative and quantitative experiments show that our method has achieved better visual quality and fidelity than the baseline on facial attribute transfer.

Original languageEnglish
Title of host publicationDigital TV and Wireless Multimedia Communication - 16th International Forum, IFTC 2019, Revised Selected Papers
EditorsGuangtao Zhai, Jun Zhou, Hua Yang, Ping An, Xiaokang Yang
PublisherSpringer Science and Business Media Deutschland GmbH
Pages91-102
Number of pages12
ISBN (Print)9789811533402
DOIs
Publication statusPublished - 2020
Externally publishedYes
Event16th International Forum on Digital TV and Wireless Multimedia Communication, IFTC 2019 - Shanghai, China
Duration: 19 Sept 201920 Sept 2019

Publication series

NameCommunications in Computer and Information Science
Volume1181
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

Conference16th International Forum on Digital TV and Wireless Multimedia Communication, IFTC 2019
Country/TerritoryChina
CityShanghai
Period19/09/1920/09/19

Keywords

  • Facial attribute transfer
  • Generative Adversarial Network
  • Multi-scale feature fusion

Fingerprint

Dive into the research topics of 'Multi-scale Generative Adversarial Learning for Facial Attribute Transfer'. Together they form a unique fingerprint.

Cite this