An Effective and Comprehensive Image Super Resolution Algorithm Combined With a Novel Convolutional Neural Network and Wavelet Transform

Abstract

In order to further improve the reconstruction effect of the image super resolution algorithm, this paper proposes an image super resolution algorithm combining deep learning and wavelet transform (ISRDW). In terms of network design, it is not only simple in structure, but also more effective in capturing image details compared with other neural network structures. At the same time, cross-connection and residual learning methods are used to reduce the difficulty of the training model. In terms of loss function, this paper uses the loss generated in the original image space domain and the wavelet domain to strengthen the constraint of network training. Experimental results show that the algorithm proposed in this paper achieves better results under different data sets and different evaluation indexes.

MFPWTN: a multi-frequency parallel wavelet transform network for remote sensing image super-resolution

“…Dou et al proposed the discrete wavelet transform SR, which also combines the residual learning with the wavelet domain to perform SR 24 and achieves satisfactory results for remote sensing images. Yang et al proposed the ISRDW 16 by combining the cross-connection and the residual learning structure and applies both the space domain loss and the wavelet domain loss to establish the total loss function.…”

“…Yang et al. proposed the ISRDW 16 by combining the cross-connection and the residual learning structure and applies both the space domain loss and the wavelet domain loss to establish the total loss function.…”

“…Hence, many studies are devoted to designing proprietary SR methods for remote sensing images, such as local-global combined network (LGCNet), 11 residual dense backprojection networks (RDBPNs), 12 and dual-parallel lightweight residual attention network (DPLRAN) 13 . For better capturing the high-frequency information and reducing the parameters of the CNN, most of studies take the wavelet transform into SR methods to propose wavelet-based SR methods, such as deep wavelet super-resolution (DWSR), 14 continuous wavelet transform and convolutional neural network (CWT-CNN), 15 image super resolution algorithm combining deep learning and wavelet transform (ISRDW), 16 and wavelet transform combined with the recursive Res-Net (WTCRR) 17 . By employing wavelet decomposition, the original image is divided into four distinct wavelet components, including three high-frequency wavelet components and one low-frequency wavelet component.…”

Multi-stage image denoising with the wavelet transform

“…To mine more useful information, wavelet transform is used into a CNN to learn detailed information and content information for image super-resolution [12]. As an alternative, using cross-connection and residual technique to integrate a CNN and wavelet transform was a good tool for image super-resolution [34]. According to mentioned illustrations, we can see that wavelet transform is effective for low-level task.…”

An Image Deblurring Method Using Improved U-Net Model

“…Highfrequency sub-bands profit to restore image texture details and constrain reconstruction of low-frequency sub-bands. Consequently, wavelet transform may be used for image processing goals such as super-resolution [24], image reconstruction [25], image defogging [26], and image deblurring [27]. In these processes, the combination of wavelet transforms and deep convolutional neural network is realized to eliminate any image blur influenced by wavelet transform, further removing blur from a noisy image.…”