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

Method for rapid super-resolution reconstruction of single image based on non-linear prediction sparse coding

A technology of super-resolution reconstruction and nonlinear prediction, which can be used in image coding, image data processing, graphics and image conversion, etc., and can solve problems such as high computational cost.

Inactive Publication Date: 2016-09-28
NANJING UNIV OF INFORMATION SCI & TECH
View PDF4 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Since the method based on sparse coding needs to solve a sparse coefficient problem for each image block in the test phase, if an image has too many blocks, the calculation overhead of the whole process is very large.

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
  • Method for rapid super-resolution reconstruction of single image based on non-linear prediction sparse coding
  • Method for rapid super-resolution reconstruction of single image based on non-linear prediction sparse coding
  • Method for rapid super-resolution reconstruction of single image based on non-linear prediction sparse coding

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0052] Attached below figure 1 The present invention is described in further detail.

[0053] In this paper, 69 images commonly used in super-resolution experiments are used as the training set, and the entire training image blocks are about 100,000 blocks. The training process is as follows:

[0054] 1) Make t=0, use Gaussian random matrix to dictionary D l and D h Initialize, and normalize each column of the dictionary as a unit, and initialize W and B randomly;

[0055] 2) fixed W (t) , B (t) , using the ADMM method to update A (t) :

[0056]

[0057] 3) Fixed A (t) , and Using the gradient descent method, update W and B:

[0058]

[0059] 4) Fixed A (t) , W (t) ,B (t) ,renew and

[0060]

[0061]

[0062] This step is optimized by using the joint dictionary training idea;

[0063] 5) make t:=t+1, iteration 2) to 4), until convergence;

[0064] After training, save D h 、D l , W, B. After obtaining the dictionary pair and model paramet...

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 nonlinear predictive sparse coding. In the training process, the present invention superimposes nonlinear predictive coding items on the basis of the classic criterion function based on sparse coding algorithms, and designs own optimization strategy to minimize the objective function, in the reconstruction process, the present invention only uses a non-linear iterative step for the input low-resolution image block and the pre-trained low-resolution dictionary to directly approximate the required sparse coding , avoiding the problem of solving sparse representation coefficients for each small image block. The experimental results show that, compared with the classical image super-resolution algorithm based on sparse coding, the present invention greatly reduces the time required for the experiment while ensuring that the reconstruction result is fully competitive.

Description

technical field [0001] The invention relates to the technical field of image information processing, in particular to a fast super-resolution reconstruction method for a single image based on predictive sparse coding. Background technique [0002] Image super-resolution has always been a basic problem in the field of digital image processing, and it is also one of the hot issues in the field of computer vision. Image super-resolution refers to the process of restoring a high-resolution image from a low-resolution image or image sequence. Since the image resolution often represents the image quality, and a high-resolution image can provide more image information, which plays a very important role in image recognition and image understanding, so recovery from low-resolution images High resolution images are very necessary. Due to the limitations of hardware equipment, the use of algorithms to realize image super-resolution technology can alleviate the complex design process ...

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
IPC IPC(8): G06T9/00G06T3/40
CPCG06T9/004G06T3/4053
Inventor 沈辉袁晓彤
Owner NANJING UNIV OF INFORMATION SCI & TECH
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