Image Super-Resolution Based on CNN Using Multilabel Gene Expression Programming

Current mainstream super-resolution algorithms based on deep learning use a deep convolution neural network (CNN) framework to realize end-to-end learning from low-resolution (LR) image to high-resolution (HR) images, and have achieved good image restoration effects. However, as the number of layers in the network is increased, better results are not necessarily obtained, and there will be problems such as slow training convergence, mismatched sample blocks, and unstable image restoration results. We propose a preclassified deep-learning algorithm (MGEP-SRCNN) using Multilabel Gene Expression Programming (MGEP), which screens out a sample sub-bank with high relevance to the target image before image block extraction, preclassifies samples in a multilabel framework, and then performs nonlinear mapping and image reconstruction. The algorithm is verified through standard images, and better objective image quality is obtained. The restoration effect under different magnification conditions is also better.

Download Full-text

Digital object restoration using generalized regression neural network deep learning—Taking Dunhuang mural restoration as an example

International Journal of Electrical Engineering Education ◽

10.1177/0020720920928549 ◽

2020 ◽

pp. 002072092092854

Author(s):

Wenjing She

Keyword(s):

Neural Network ◽

Deep Learning ◽

Image Restoration ◽

Anisotropic Diffusion ◽

Learning Algorithm ◽

Diffusion Method ◽

Generalized Regression Neural Network ◽

Deep Learning Algorithm ◽

Training Samples ◽

Restoration Effect

In this research, Dunhuang murals is taken as the object of restoration, and the role of digital repair combined with deep learning algorithm in mural restoration is explored. First, the image restoration technology is described, as well as its advantages and disadvantages are analyzed. Second, the deep learning algorithm based on artificial neural network is described and analyzed. Finally, the deep learning algorithm is integrated into the digital repair technology, and a mural restoration method based on the generalized regression neural network is proposed. The morphological expansion method and anisotropic diffusion method are used to preprocess the image. The MATLAB software is used for the simulation analysis and evaluation of the image restoration effect. The results show that in the restoration of the original image, the accuracy of the digital image restoration technology is not high. The nontexture restoration technology is not applicable in the repair of large-scale texture areas. The predicted value of the mural restoration effect based on the generalized neural network is closer to the true value. The anisotropic diffusion method has a significant effect on the processing of image noise. In the image similarity rate, the different number of training samples and smoothing parameters are compared and analyzed. It is found that when the value of δ is small, the number of training samples should be increased to improve the accuracy of the prediction value. If the number of training samples is small, a larger value of δ is needed to get a better prediction effect, and the best restoration effect is obtained for the restored image. Through this study, it is found that this study has a good effect on the restoration model of Dunhuang murals. It provides experimental reference for the restoration of later murals.

Download Full-text

Images super-resolution by optimal deep AlexNet architecture for medical application: A novel DOCALN1

Journal of Intelligent & Fuzzy Systems ◽

10.3233/jifs-189146 ◽

2020 ◽

Vol 39 (6) ◽

pp. 8259-8272

Author(s):

Sudhakar Sengan ◽

L. Arokia Jesu Prabhu ◽

V. Ramachandran ◽

V. Priya ◽

Logesh Ravi ◽

...

Keyword(s):

Deep Learning ◽

Image Quality Assessment ◽

Learning Algorithm ◽

Super Resolution ◽

Medical Application ◽

Activation Function ◽

Convolutional Network ◽

Proposed Model ◽

Image Super Resolution ◽

Optimal Values

In the last decade, numerous researches have been focused on Image Super-Resolution (SR); this recreation or improvement model is vital in different research areas. Recently, deep learning algorithm finds useful to advance in the resolution of the medical output. Here, we devise a novel Deep Convolutional Network model along with the optimal learning rate of the Rectified Linear Unit (ReLU) intended for Medical Image Super-Resolution (MISR). For getting the optimal values of Deep Learning AlexNet structure, Modified Crow Search (MCS) is utilized, which is mainly depends on the behavior of crow sets. The chosen Alexnet lacks in a sort of suitable supervision for upgrading execution of the proposed model that effectively aims to overfit. The proposed design, i.e., MISR, named Deep Optimal Convolutional AlexNet (DOCALN), derives the optimal values of learning rates of the ReLU activation function. Based on this optimal deep learning structure, the Low Resolution (LR) medical images can be applied. Experimentation results of our proposed model are compared with variants of Convolution Neural Networks (CNN) concerning different measures such as image quality assessment, SR efficiency analysis, and execution time.

Download Full-text

Image Restoration Based on Adaptive Dual-Domain Filtering

Mathematical Problems in Engineering ◽

10.1155/2018/4790174 ◽

2018 ◽

Vol 2018 ◽

pp. 1-17

Author(s):

Ruiqiang He ◽

Xiangchu Feng ◽

Chenping Zhao ◽

Huazhu Chen ◽

Xiaolong Zhu ◽

...

Keyword(s):

Image Restoration ◽

Learning Algorithm ◽

Super Resolution ◽

Image Deblurring ◽

Superior Performance ◽

Single Image ◽

Denoising Method ◽

Image Super Resolution ◽

Single Image Super Resolution ◽

Dual Domain

Image restoration is a long-standing problem in low-level computer vision. In this paper, we offer a simple but effective estimation paradigm for various image restoration problems. Specifically, we first propose a model-based Gaussian denoising method Adaptive Dual-Domain Filtering (ADDF) by learning the optimal confidence factors which are adjusted adaptively with Gaussian noise standard deviation. In addition, by generalizing this learning approach to Laplace noise, the learning algorithm of the optimum confidence factors in Laplace denoising is presented. Finally, the proposed ADDF is tactfully plugged into the method frameworks of off-the-shelf image deblurring and single image super-resolution (SISR). The approach, coining the name Plug-ADDF, achieves promising performance. Extensive experiments validate that the proposed ADDF for Gaussian and Laplace noise removals indeed results in visual and quantitative improvements over some existing state-of-the-art methods. Moreover, our Plug-ADDF for image deblurring and SISR also demonstrates superior performance objectively and subjectively.

Download Full-text

Single Target SAR 3D Reconstruction Based on Deep Learning

Sensors ◽

10.3390/s21030964 ◽

2021 ◽

Vol 21 (3) ◽

pp. 964

Author(s):

Shihong Wang ◽

Jiayi Guo ◽

Yueting Zhang ◽

Yuxin Hu ◽

Chibiao Ding ◽

...

Keyword(s):

Deep Learning ◽

3D Reconstruction ◽

Time Window ◽

Learning Algorithm ◽

Signal To Noise Ratio ◽

Super Resolution ◽

Mapping Method ◽

Deep Learning Algorithm ◽

Deep Learning Network ◽

Voxel Grid

Synthetic aperture radar tomography (TomoSAR) is an important 3D mapping method. Traditional TomoSAR requires a large number of observation orbits however, it is hard to meet the requirement of massive orbits. While on the one hand, this is due to funding constraints, on the other hand, because the target scene is changing over time and each observation orbit consumes lots of time, the number of orbits can be fewer as required within a narrow time window. When the number of observation orbits is insufficient, the signal-to-noise ratio (SNR), peak-to-sidelobe ratio (PSR), and resolution of 3D reconstruction results will decline severely, which seriously limits the practical application of TomoSAR. In order to solve this problem, we propose to use a deep learning network to improve the resolution and SNR of 3D reconstruction results under the condition of very few observation orbits by learning the prior distribution of targets. We use all available orbits to reconstruct a high resolution target, while only very few (around 3) orbits to reconstruct a low resolution input. The low-res and high-res 3D voxel-grid pairs are used to train a 3D super-resolution (SR) CNN (convolutional neural network) model, just like ordinary 2D image SR tasks. Experiments on the Civilian Vehicle Radar dataset show that the proposed deep learning algorithm can effectively improve the reconstruction both in quality and in quantity. In addition, the model also shows good generalization performance for targets not shown in the training set.

Download Full-text