An Omnidirectional Image Super-Resolution Method Based on Enhanced SwinIR
Open Access
- 28 April 2024
- journal article
- research article
- Published by MDPI AG in Information
- Vol. 15 (5) , 248
Abstract
For the significant distortion problem caused by the special projection method of equi-rectangular projection (ERP) images, this paper proposes an omnidirectional image super-resolution algorithm model based on position information transformation, taking SwinIR as the base. By introducing a space position transformation module that supports deformable convolution, the image preprocessing process is optimized to reduce the distortion effects in the polar regions of the ERP image. Meanwhile, by introducing deformable convolution in the deep feature extraction process, the model’s adaptability to local deformations of images is enhanced. Experimental results on publicly available datasets have shown that our method outperforms SwinIR, with an average improvement of over 0.2 dB in WS-PSNR and over 0.030 in WS-SSIM for ×4 pixel upscaling.Keywords
Funding Information
- National Natural Science Foundation of China (61972050, 62201525, 62272040, 62172005, JYB2022-01)
- Research on the strategic project of the science and Technology Commission of the Ministry of Education of China (61972050, 62201525, 62272040, 62172005, JYB2022-01)
This publication has 32 references indexed in Scilit:
- Metric Learning Based Interactive Modulation for Real-World Super-Resolution2022
- ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks2019
- Image Super-Resolution Using Very Deep Residual Channel Attention Networks2018
- Weighted-to-Spherically-Uniform Quality Evaluation for Omnidirectional VideoIEEE Signal Processing Letters, 2017
- Deeply-Recursive Convolutional Network for Image Super-Resolution2016
- Deep Residual Learning for Image Recognition2016
- Accurate Image Super-Resolution Using Very Deep Convolutional Networks2016
- Image Super-Resolution Using Deep Convolutional NetworksIEEE Transactions on Pattern Analysis and Machine Intelligence, 2015
- Single Image Super-Resolution With Non-Local Means and Steering Kernel RegressionIEEE Transactions on Image Processing, 2012
- Low-Complexity Single-Image Super-Resolution based on Nonnegative Neighbor Embedding2012