A Review of Single Image Super-Resolution Reconstruction Based on Deep Learning

被引：0
作者：
Wu J. [1 ,2 ]
Ye X.-J. [1 ,2 ]
Huang F. [1 ,2 ]
Chen L.-Q. [1 ,2 ]
Wang Z.-F. [1 ,2 ]
Liu W.-X. [2 ,3 ]
机构：
[1] School of Mechanical Engineering and Automation, Fuzhou University, Fujian, Fuzhou
[2] Advanced Technology Innovation Institute, Fuzhou University, Fujian, Fuzhou
[3] College of Computer and Data Science, Fuzhou University, Fujian, Fuzhou
来源：
Tien Tzu Hsueh Pao/Acta Electronica Sinica | 2022年 / 50卷 / 09期
关键词：
challenges; convolutional neural network; deep learning; generative adversarial network; single image; super-resolution reconstruction; transformer;
D O I：
10.12263/DZXB.20220091
中图分类号：
学科分类号：
摘要：
Image super-resolution reconstruction is one of the basic image processing techniques in computer vision, which can not only improve image resolution and image quality, but also assist other computer vision tasks. In recent years, with the rise of artificial intelligence, deep-learning-based image super-resolution reconstruction has also made remarkable progress. Based on a brief description of the image super-resolution reconstruction methodology, this paper comprehensive⁃ ly reviews the technical architecture and research process of deep-learning-based single image super-resolution reconstruc⁃ tion, including the method of datasets construction, the basic framework of the network model, the subjective and objective evaluation metrics for image quality evaluation. The methods based on convolutional neural networks, generative adversari⁃ al networks and Transformer, which are divided according to network structure and image reconstruction effect are mainly introduced, and related network models are reviewed and compared. Finally, the future development trend of image super-resolution reconstruction is prospected according to the related content of network model and super-resolution reconstruc⁃ tion challenges. © 2022 Chinese Institute of Electronics. All rights reserved.
引用
收藏
页码：2265 / 2294
页数：29
相关论文
共 158 条
[71]  
KIM J, LEE J K, LEE K M., Deeply-recursive convolu⁃ tional network for image super-resolution, 2016 IEEE Conference on Computer Vision and Pattern Recognition, pp. 1637-1645, (2016)
[72]  
HU X C, MU H Y, ZHANG X Y, Et al., Meta-SR: A mag⁃ nification-arbitrary network for super-resolution, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR), pp. 1575-1584, (2019)
[73]  
SHI W Z, CABALLERO J, HUSZAR F, Et al., Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network, 2016 IEEE Conference on Computer Vision and Pattern Recognition, pp. 1874-1883, (2016)
[74]  
FAN Y C, SHI H H, YU J H, Et al., Balanced two-stage re⁃ sidual networks for image super-resolution, 2017 IEEE Conference on Computer Vision and Pattern Recog⁃ nition Workshops, pp. 1157-1164, (2017)
[75]  
ZHANG Y, LI K, LI K, Et al., Residual non-local attention networks for image restoration, Proceedings of the Seventh International Conference on Learning Represen⁃ tations(ICLR 2019), pp. 1-18, (2019)
[76]  
LAI W S, HUANG J B, AHUJA N, Et al., Deep laplacian pyramid networks for fast and accurate super-resolution, 2017 IEEE Conference on Computer Vision and Pat⁃ tern Recognition(CVPR), pp. 5835-5843, (2017)
[77]  
LAI W S, HUANG J B, AHUJA N, Et al., Fast and accu⁃ rate image super-resolution with deep Laplacian pyramid networks, IEEE Transactions on Pattern Analysis and Machine Intelligence, 41, 11, pp. 2599-2613, (2019)
[78]  
CHUDASAMA V, UPLA K., E-ProSRNet: An enhanced progressive single image super-resolution approach, Computer Vision and Image Understanding, 200, (2020)
[79]  
HARIS M, SHAKHNAROVICH G, UKITA N., Deep back-projection networks for super-resolution, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 1664-1673, (2018)
[80]  
HARIS M, SHAKHNAROVICH G, UKITA N., Deep back-Projection Networks for single image super-resolu⁃ tion, IEEE Transactions on Pattern Analysis and Ma⁃ chine Intelligence, 43, 12, pp. 4323-4337, (2021)
3456789101112