Inversion Based on a Detached Dual-Channel Domain Method for StyleGAN2 Embedding

Authors

Nan Yang, Chinese Academy of Sciences
Mengchu Zhou, Newark College of Engineering
Bingjie Xia, Chinese Academy of Sciences
Xiwang Guo, Newark College of Engineering
Liang Qi, Shandong University of Science and Technology

Document Type

Article

Publication Date

1-1-2021

Abstract

A style-based generative adversarial network (StyleGAN2) yields remarkable results in image-to-latent embedding. This work proposes a Detached Dual-channel Domain Encoder as an effective and robust method to embed an image to a latent code, i.e., GAN inversion. It infers a latent code from two aspects: a) a detached dual-channel design to support faithful image reconstruction; and b) a local skip connection that allows conveying pieces of information with image details. We further introduce a hierarchical progressive training strategy that allows the proposed encoder to separately capture different semantic features. The qualitative and quantitative experimental results show that the well-trained encoder can embed an image into a latent code in StyleGAN2 latent space with less time than its peers while preserving facial identity and image details well.

Identifier

85102238360 (Scopus)

Publication Title

IEEE Signal Processing Letters

External Full Text Location

https://doi.org/10.1109/LSP.2021.3059371

e-ISSN

15582361

ISSN

10709908

First Page

553

Last Page

557

Volume

28

Grant

61773367

Fund Ref

National Natural Science Foundation of China

Recommended Citation

Yang, Nan; Zhou, Mengchu; Xia, Bingjie; Guo, Xiwang; and Qi, Liang, "Inversion Based on a Detached Dual-Channel Domain Method for StyleGAN2 Embedding" (2021). Faculty Publications. 4703.
https://digitalcommons.njit.edu/fac_pubs/4703