Recent Developments on Super-Resolution: A Comprehensive Overview

Super-resolution technology aims to reconstruct high-resolution images from low-resolution inputs, with applications in video surveillance, medical diagnosis, and remote sensing. Various convolutional neural network (CNN) models have been developed, such as SRCNN, VDSR, ESPCN, and FSRCNN, each with its unique approaches to enhance image resolution. Techniques like deeper network structures, larger receptive fields, and efficient sub-pixel convolution have significantly improved the quality of super-resolved images.

• Super-resolution
• Deep learning
• Convolutional neural network
• Image processing
• High-resolution

donnie Follow

Uploaded on Aug 11, 2024 | 0 Views

Download Presentation

Please find below an Image/Link to download the presentation.

The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. Download presentation by click this link. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.

Presentation Transcript

1. Recent Developments on Super-Resolution Yi-Wen Chen 2018.12.28

2. 2 Super-Resolution Goal Reconstruct a high-resolution (HR) image from a low-resolution (LR) input image Application Video surveillance Medical diagnosis Remote sensing

3. 3 CNN-based SR SRCNN [Dong et al. ECCV 14] first CNN-based SR VDSR [Kim et al. CVPR 16] increase the network depth ESPCN [Shi et al. CVPR 16] bicubic upsampling sub-pixel convolution FSRCNN [Dong et al. ECCV 16] bicubic upsampling transposed convolution

4. 4 Ground Truth SRCNN Bicubic interpolation Convolutional layer C. Dong, C. C. Loy, K. He, and X. Tang. Learning a deep convolutional network for image super-resolution. In ECCV, 2014.

5. 5 SRCNN VDSR Context deeper network (3 layers 20 layers) larger receptive field Convergence higher learning rate enabled by residual-learning and gradient clipping (training time: 3 days 4 hours) Scale Factor cope with multiscale SR problem in a single network

6. 6 VDSR Bicubic interpolation J. Kim, J. K. Lee, and K. M. Lee. Accurate image super-resolution using very deep convolutional networks. In CVPR, 2016.

7. 7 VDSR J. Kim, J. K. Lee, and K. M. Lee. Accurate image super-resolution using very deep convolutional networks. In CVPR, 2016.

8. 8 ESPCN W. Shi, J. Caballero, F. Huszar, J. Totz, A. Aitken, R. Bishop, D. Rueckert, and Z. Wang. Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network. In CVPR, 2016.

9. 9 ESPCN W. Shi, J. Caballero, F. Huszar, J. Totz, A. Aitken, R. Bishop, D. Rueckert, and Z. Wang. Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network. In CVPR, 2016.

10. 10 FSRCNN C. Dong, C. C. Loy, and X. Tang. Accelerating the super-resolution convolutional neural network. In ECCV, 2016.

11. 11 FSRCNN C. Dong, C. C. Loy, and X. Tang. Accelerating the super-resolution convolutional neural network. In ECCV, 2016.

12. 12 References C. Dong, C. C. Loy, K. He, and X. Tang. Learning a deep convolutional network for image super-resolution. In ECCV, 2014. J. Kim, J. K. Lee, and K. M. Lee. Accurate image super- resolution using very deep convolutional networks. In CVPR, 2016. W. Shi, J. Caballero, F. Huszar, J. Totz, A. Aitken, R. Bishop, D. Rueckert, and Z. Wang. Real-time single image and video super-resolution using an efficient sub- pixel convolutional neural network. In CVPR, 2016. C. Dong, C. C. Loy, and X. Tang. Accelerating the super- resolution convolutional neural network. In ECCV, 2016.