Cascaded face super-resolution with shape and identity priors

Zeng, Dan1,2; Li, Zelin1,2; Yan, Xiao1,2; Jiang, Wen1,2; Wang, Xinshao3; Liu, Jiang1,2; Tang, Bo1,2

2023-07-01

10.1049/ipr2.12863

IET IMAGE PROCESSING   影响因子和分区

1751-9659

1751-9667

Despite impressive progress in face super-resolution (SR), it is an open challenge to reconstruct a reliable SR face that preserves authentic facial characteristics. Here, the problem of super-resolving low-resolution (LR) faces to high-resolution (HR) ones is addressed. To tackle the ill-posed nature of face SR, the cascaded super-resolution network (CSRNet) is proposed to utilize shape and identity priors jointly and progressively, the first to explore multiple priors. Specifically, CSRNet adopts a cascaded structure to transform an LR face to HR face progressively via multiple stages. At each stage, CSRNet forces its output face image to match both the shape priors and identity priors extracted from the ground-truth HR face. The shape priors estimated in one stage are merged into the inputs of its subsequent stage to provide rich information for the face SR. To generate realistic yet discriminative faces, the cascaded super-resolution generative adversarial network (CSRGAN) is also proposed to incorporate the adversarial loss and identification loss into CSRNet. Extensive experiments on popular benchmarks show that the CSRNet and CSRGAN outperform existing face SR state-of-the-art methods, both quantitatively and qualitatively, and detailed ablation studies show the advantage of this method.

image enhancement image reconstruction image resolution

SCI ; EI

英语

第一 ; 通讯

National Natural Science Foundation of China[62206123]

Computer Science ; Engineering ; Imaging Science & Photographic Technology

Computer Science, Artificial Intelligence ; Engineering, Electrical & Electronic ; Imaging Science & Photographic Technology

WOS:001021149600001

WILEY

20232914397559

Generative adversarial networks ; Image reconstruction ; Image resolution

Artificial Intelligence:723.4

Web of Science

1.Southern Univ Sci & Technol, Res Inst Trustworthy Autonomous Syst, Shenzhen 518055, Peoples R China
2.Southern Univ Sci & Technol, Dept Comp Sci & Engn, Shenzhen 518055, Peoples R China
3.Shenzhen Univ, Coll Comp Sci & Software Engn, Shenzhen, Peoples R China

Zeng, Dan,Li, Zelin,Yan, Xiao,et al. Cascaded face super-resolution with shape and identity priors[J]. IET IMAGE PROCESSING,2023,17(11):3309-3322.

Zeng, Dan.,Li, Zelin.,Yan, Xiao.,Jiang, Wen.,Wang, Xinshao.,...&Tang, Bo.(2023).Cascaded face super-resolution with shape and identity priors.IET IMAGE PROCESSING,17(11),3309-3322.

Zeng, Dan,et al."Cascaded face super-resolution with shape and identity priors".IET IMAGE PROCESSING 17.11(2023):3309-3322.