High-quality face image super-resolution based on Generative Adversarial Networks

Xinru Zhong; Xiujie Qu; Chen Chen

doi:10.1109/IAEAC47372.2019.8998075

High-quality face image super-resolution based on Generative Adversarial Networks

Xinru Zhong, Xiujie Qu, Chen Chen

Beijing Institute of Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

4 Citations (Scopus)

Abstract

Face image super-resolution has received increasing attention. However, since the face has a lot of fine textures, it is very difficult to rebuild for large upscaling factors. We propose a new method for face image SR, using residul dense block(RDB) as the basic unit and the Inception architecture is combined in the low layers. We use the relativistic GAN and the improved perceptual loss defined by the features before activation.For the large scaling factors, our GAN is progressive both in architecture and training. The network proposed achieves excellent performance in the reconstruction of low-resolution face images, especially under large scaling factors such as 4x and 8x.

Original language	English
Title of host publication	Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019
Editors	Bing Xu, Kefen Mou
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	1178-1182
Number of pages	5
ISBN (Electronic)	9781728119076
DOIs	https://doi.org/10.1109/IAEAC47372.2019.8998075
Publication status	Published - Dec 2019
Event	4th IEEE Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019 - Chengdu, China Duration: 20 Dec 2019 → 22 Dec 2019

Publication series

Name	Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019

Conference

Conference	4th IEEE Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019
Country/Territory	China
City	Chengdu
Period	20/12/19 → 22/12/19

Keywords

face image super-resolution
progressive trainging
relativistic Generative Adversarial Networks

Access to Document

10.1109/IAEAC47372.2019.8998075

Cite this

Zhong, X., Qu, X., & Chen, C. (2019). High-quality face image super-resolution based on Generative Adversarial Networks. In B. Xu, & K. Mou (Eds.), Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019 (pp. 1178-1182). Article 8998075 (Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IAEAC47372.2019.8998075

Zhong, Xinru ; Qu, Xiujie ; Chen, Chen. / High-quality face image super-resolution based on Generative Adversarial Networks. Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019. editor / Bing Xu ; Kefen Mou. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 1178-1182 (Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019).

@inproceedings{04d5315343424bc8a95cb7b3d74150dc,

title = "High-quality face image super-resolution based on Generative Adversarial Networks",

abstract = "Face image super-resolution has received increasing attention. However, since the face has a lot of fine textures, it is very difficult to rebuild for large upscaling factors. We propose a new method for face image SR, using residul dense block(RDB) as the basic unit and the Inception architecture is combined in the low layers. We use the relativistic GAN and the improved perceptual loss defined by the features before activation.For the large scaling factors, our GAN is progressive both in architecture and training. The network proposed achieves excellent performance in the reconstruction of low-resolution face images, especially under large scaling factors such as 4x and 8x.",

keywords = "face image super-resolution, progressive trainging, relativistic Generative Adversarial Networks",

author = "Xinru Zhong and Xiujie Qu and Chen Chen",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 4th IEEE Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019 ; Conference date: 20-12-2019 Through 22-12-2019",

year = "2019",

month = dec,

doi = "10.1109/IAEAC47372.2019.8998075",

language = "English",

series = "Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "1178--1182",

editor = "Bing Xu and Kefen Mou",

booktitle = "Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019",

address = "United States",

}

Zhong, X, Qu, X & Chen, C 2019, High-quality face image super-resolution based on Generative Adversarial Networks. in B Xu & K Mou (eds), Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019., 8998075, Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019, Institute of Electrical and Electronics Engineers Inc., pp. 1178-1182, 4th IEEE Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019, Chengdu, China, 20/12/19. https://doi.org/10.1109/IAEAC47372.2019.8998075

High-quality face image super-resolution based on Generative Adversarial Networks. / Zhong, Xinru; Qu, Xiujie ; Chen, Chen.
Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019. ed. / Bing Xu; Kefen Mou. Institute of Electrical and Electronics Engineers Inc., 2019. p. 1178-1182 8998075 (Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - High-quality face image super-resolution based on Generative Adversarial Networks

AU - Zhong, Xinru

AU - Qu, Xiujie

AU - Chen, Chen

PY - 2019/12

Y1 - 2019/12

N2 - Face image super-resolution has received increasing attention. However, since the face has a lot of fine textures, it is very difficult to rebuild for large upscaling factors. We propose a new method for face image SR, using residul dense block(RDB) as the basic unit and the Inception architecture is combined in the low layers. We use the relativistic GAN and the improved perceptual loss defined by the features before activation.For the large scaling factors, our GAN is progressive both in architecture and training. The network proposed achieves excellent performance in the reconstruction of low-resolution face images, especially under large scaling factors such as 4x and 8x.

AB - Face image super-resolution has received increasing attention. However, since the face has a lot of fine textures, it is very difficult to rebuild for large upscaling factors. We propose a new method for face image SR, using residul dense block(RDB) as the basic unit and the Inception architecture is combined in the low layers. We use the relativistic GAN and the improved perceptual loss defined by the features before activation.For the large scaling factors, our GAN is progressive both in architecture and training. The network proposed achieves excellent performance in the reconstruction of low-resolution face images, especially under large scaling factors such as 4x and 8x.

KW - face image super-resolution

KW - progressive trainging

KW - relativistic Generative Adversarial Networks

UR - http://www.scopus.com/inward/record.url?scp=85081172504&partnerID=8YFLogxK

U2 - 10.1109/IAEAC47372.2019.8998075

DO - 10.1109/IAEAC47372.2019.8998075

M3 - Conference contribution

AN - SCOPUS:85081172504

T3 - Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019

SP - 1178

EP - 1182

BT - Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019

A2 - Xu, Bing

A2 - Mou, Kefen

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 4th IEEE Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019

Y2 - 20 December 2019 through 22 December 2019

ER -

Zhong X, Qu X , Chen C. High-quality face image super-resolution based on Generative Adversarial Networks. In Xu B, Mou K, editors, Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019. Institute of Electrical and Electronics Engineers Inc. 2019. p. 1178-1182. 8998075. (Proceedings of 2019 IEEE 4th Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2019). doi: 10.1109/IAEAC47372.2019.8998075

High-quality face image super-resolution based on Generative Adversarial Networks

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this