Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

An image super-resolution reconstruction method based on sparse representation and deep learning

A technology of super-resolution reconstruction and sparse representation, applied in the field of image processing, can solve the problems of time, high computational complexity, poor performance of super-resolution images, poor performance of restored images, etc. Effect

Active Publication Date: 2019-05-10
XIDIAN UNIV
View PDF3 Cites 50 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004]To sum up: the high-frequency detail information of the high-resolution image reconstructed based on the traditional sparse representation is not rich enough, and the image restoration performance is poor; in the original deep learning network structure, time, The computational complexity is large, and the residual network restores super-resolution images through bicubic interpolation with poor performance

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • An image super-resolution reconstruction method based on sparse representation and deep learning
  • An image super-resolution reconstruction method based on sparse representation and deep learning
  • An image super-resolution reconstruction method based on sparse representation and deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0027] Images contain more information than text, and can be intuitively provided to humans. In applications such as remote sensing monitoring, military reconnaissance, and transportation, image super-resolution reconstruction provides rich detailed information. However, based on traditional The high-frequency detail information of the high-resolution image reconstructed by the sparse representation of the sparse representation is not rich enough, has edge effects, and the performance of restoring the image is poor; the original deep learning network structure is huge, and the number of layers is complex, which leads to large time and computational complexity, and the residual image based on deep learning Poor network recovers super-resolution images via bicubic interpolation with poor performance. In response to these problems, the present invention conducts research and proposes an image super-resolution reconstruction method based on sparse representation and deep learning, ...

Embodiment 2

[0037] The image super-resolution reconstruction method based on sparse representation and deep learning is the same as in embodiment 1, and the sparse representation theory and dictionary learning in step 2, the sparse coefficient describes the following mathematical expression:

[0038] or

[0039] where x∈R m is a vector representation of an image (block), D∈R N*L is an over-complete dictionary, a is the sparse coefficient of image x under dictionary D, if only a few coefficients in a are not zero, ie ||a|| 0 <<N, the image x is sparse.

[0040] On the basis of the above sparsity theory, solve all the training set constraint optimization problems, and obtain the sparse matrix α and the over-complete dictionary D:

[0041]

[0042] Among them, X is the training sample pair, D h and D l is a high-resolution dictionary and a low-resolution dictionary with a consistent sparse matrix α, and λ is a regularization parameter;

[0043] 1. Initialization: D (0) ∈ R n×K...

Embodiment 3

[0053] The image super-resolution reconstruction method based on sparse representation and deep learning is the same as embodiment 1-2, the dictionary joint training described in step 2, and the high-resolution dictionary D is obtained respectively through traditional sparse representation theoretical training. h and a low-resolution dictionary D l , in the dictionary training phase, in order to ensure that the same sparse coefficients can be obtained, the training is performed according to the following formula:

[0054]

[0055]

[0056] In order to meet the isomorphism of the two dictionaries under the sparse representation and the difference of different channels, and to ensure that the sparse matrix is ​​the same, combining the above two formulas needs to solve the minimum constraint formula:

[0057]

[0058] Among them, i represents the three channels of y, cb, and cr, that is, the YUV brightness and chrominance channels, and N and M represent the vector dimens...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses an image super-resolution reconstruction method based on sparse representation and deep learning, and solves the problems that the image super-resolution process is complex incalculation and the quality of a reconstructed image is poor. The method comprises the following implementation steps: collecting and extracting training data blocks and a chromaticity and brightnessdictionary for combined optimization training; independently reconstructing a high-resolution image block; Carrying out high-resolution image reconstruction of sparse representation; Training a residual error network based on deep learning to optimize high-frequency details; Image super-resolution reconstruction. In order to prevent an edge effect and a fuzzy effect, chroma and brightness data aredistinguished and independently reconstructed; in order to optimize high-frequency detail information of a sparse representation output high-resolution image, the high-resolution image based on sparse representation reconstruction is input into a residual network, and a high-frequency residual image is output through four times of convolution feature extraction and feature fusion and input bitwise addition to reconstruct a super-resolution image. The method is low in calculation complexity, high in image reconstruction quality and widely applied to the fields of remote sensing monitoring, criminal investigation, traffic management and the like.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to super-resolution image reconstruction, in particular to an image super-resolution reconstruction method based on sparse representation and deep learning. It is used in applications such as remote sensing monitoring, military reconnaissance, traffic and security monitoring, medical diagnosis and pattern recognition, and high-definition video signals. Background technique [0002] In the process of image acquisition, due to the limitation of imaging distance, imaging device resolution and other factors, it is difficult for the imaging system to obtain the information in the original scene without distortion. The imaging system is usually affected by many factors such as deformation, blur, downsampling and noise. This results in a decrease in the quality of the acquired image. Therefore, how to improve the spatial resolution and image quality of images has alwa...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06T3/40G06T5/50
Inventor 张静于露露李云松尹雅平王卓
Owner XIDIAN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com