StarSRGAN: Improving Real-World Blind Super-Resolution

被引：0

作者：

Vo K.D. ^{[1
]}

Bui L.T. ^{[1
]}

机构：

[1] Faculty of Information Technology (FIT), University of Science, VNU.HCM, Ho Chi Minh City

来源：

Computer Science Research Notes | 2023年 / 31卷 / 1-2期

关键词：

adaptive degradation; Blind super-resolution; dropout degradation; dual perceptual loss; multi-scale discriminator;

D O I：

10.24132/CSRN.3301.9

中图分类号：

学科分类号：

摘要：

The aim of blind super-resolution (SR) in computer vision is to improve the resolution of an image without prior knowledge of the degradation process that caused the image to be low-resolution. The State of the Art (SOTA) model Real-ESRGAN has advanced perceptual loss and produced visually compelling outcomes using more complex degradation models to simulate real-world degradations. However, there is still room to improve the super-resolved quality of Real-ESRGAN by implementing recent techniques. This research paper introduces StarSRGAN, a novel GAN model designed for blind super-resolution tasks that utilize 5 various architectures. Our model provides new SOTA performance with roughly 10% better on the MANIQA and AHIQ measures, as demonstrated by experimental comparisons with Real-ESRGAN. In addition, as a compact version, StarSRGAN Lite provides approximately 7.5 times faster reconstruction speed (real-time upsampling from 540p to 4K) but can still keep nearly 90% of image quality, thereby facilitating the development of a real-time SR experience for future research. Our codes are released at https://github.com/kynthesis/StarSRGAN. © 2023 Seventh Sense Research Group®

引用

页码：62 / 72

页数：10

共 26 条

[1]

Dong C., Et al., Image Super-Resolution Using Deep Convolutional Networks, IEEE TPAMI 2016, 38, 2, pp. 295-307, (2016)

[2]

Johnson J., Et al., Perceptual Losses for Real-Time Style Transfer and Super-Resolution, ECCV 2016, pp. 694-711, (2016)

[3]

Ledig C., Et al., Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network, IEEE/CVF CVPR 2017, pp. 105-114, (2017)

[4]

Wang X., Et al., ESRGAN: Enhanced super-resolution generative adversarial networks, pp. 63-79, (2019)

[5]

Goodfellow I., Et al., Generative Adversarial Networks, NeurIPS, 2014, (2014)

[6]

Wang X., Et al., Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data, IEEE/CVF IC-CVW 2021, pp. 1905-1914, (2021)

[7]

Rakotonirina N. C., Rasoanaivo A., ESRGAN+: Further Improving Enhanced Super-Resolution Generative Adversarial Network, IEEE ICASSP 2020, pp. 3637-3641, (2020)

[8]

Wei Z., Et al., A-ESRGAN: Training Real-World Blind Super-Resolution with Attention U-Net Discriminators, (2021)

[9]

Liang J., Zeng H., Zhang L., Efficient And Degradation-Adaptive Network For Real-World Image Super-Resolution, ECCV, 2022, pp. 574-591, (2022)

[10]

Song J., Et al., Dual Perceptual Loss for Single Image Super-Resolution Using ESRGAN, (2022)

← 1 2 3 →