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

Image super-resolution reconstruction method based on representational learning and neighbor constraint embedding

A super-resolution and neighborhood technology, applied in the field of image processing, can solve the problems of general reconstructed image quality, missing neighbors, and inaccurate extraction, etc., to overcome poor compatibility, improve restoration effect, detail and texture information Good results

Active Publication Date: 2015-11-18
XIDIAN UNIV
View PDF1 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the feature extraction of these methods is not accurate enough, and accurate neighbors are often not found, and the size of the neighborhood is also fixed, the efficiency of neighborhood embedding is low, and the quality of the reconstructed image is average.

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
  • Image super-resolution reconstruction method based on representational learning and neighbor constraint embedding
  • Image super-resolution reconstruction method based on representational learning and neighbor constraint embedding
  • Image super-resolution reconstruction method based on representational learning and neighbor constraint embedding

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0030] The present invention is an image super-resolution reconstruction method based on representation learning and neighborhood constraint embedding, see figure 1 , the present invention includes the following steps to image super-resolution reconstruction:

[0031] Step 1: Input the training sample image pair, and use the low-resolution training sample image to learn to construct a low-resolution image block dictionary with a scale of 10W And use the high-resolution training sample images to learn to construct a dictionary with a size of N and low-resolution image blocks The corresponding high-resolution image patch dictionary The training image pairs used are standard natural images commonly used in the field of image processing, see figure 2 , image 3 , Figure 4 .

[0032] Step 2: Input the low-resolution test image Y to be super-resolution reconstructed, and divide it into blocks in an overlapping manner to obtain a low-resolution image block set For each bl...

Embodiment 2

[0040] The image super-resolution reconstruction method based on representation learning and neighborhood constraint embedding is the same as in Embodiment 1. Step 1 and step 2 in the present invention both involve the process of feature extraction, and both use a deep sparse autoencoder network to learn the characteristics of image blocks. Therefore, it is summarized as follows:

[0041] Wherein step 1 trains the training sample image pairs, including the following steps:

[0042] 1a) Input the training image pair, block the high-resolution and low-resolution images respectively, and obtain the low-resolution image block set and the corresponding set of high-resolution image patches The training image pairs used are standard natural images commonly used in the field of image processing, see figure 2 , image 3 , Figure 4 ;

[0043] 1b) For low-resolution training image patch set X sp , through the training of deep sparse autoencoder network to obtain cascaded featur...

Embodiment 3

[0050] The image super-resolution reconstruction method based on representation learning and neighborhood constraint embedding is the same as that in Embodiment 1-2.

[0051] Wherein the pre-selected feature block neighbor described in step 3 includes the following steps:

[0052] 3a) For an input low-resolution image feature block Compute feature blocks Dictionaries with low-resolution image patches the Euclidean distance;

[0053] 3b) Select the first K small blocks with the smallest distance as The preselected feature neighbors of yes In the low-resolution image patch dictionary X s The initial neighborhood in .

[0054] In the present invention, K=100 training low-resolution image blocks are selected as initial neighbors, and these 100 low-resolution image blocks are the most similar to the input feature block and have a fixed size. Too many neighbors will increase the complexity of the algorithm and slow down the operation speed, and too few neighbors will n...

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 representational learning and neighbor constraint embedding, and problems of inaccurate feature extraction and fixed size of neighbors are solved. The method includes main steps: pre-processing a group of training sample images, and constructing a low-resolution image block dictionary and a high-resolution image block dictionary; inputting a low-resolution test image, and performing partitioning and extracting features of the low-resolution test image; calculating the Euclidean distance between the features, and searching K neighbors of image blocks; and constructing an adaptive constraint function, obtaining k adaptive neighbors via neighbor constraint, obtaining a final high-resolution image by employing a locally linear embedding method, and accomplishing the image super-resolution reconstruction. According to the method, the characteristic of deep sparse self-coding network learning is employed, neighbor selection is accurate, the size of the neighbors is selected in an adaptive manner, the reconstruction image quality is effectively enhanced, detailed information is improved, and the method is applicable to super-resolution reconstruction of various natural images.

Description

technical field [0001] The invention belongs to the technical field of image processing, and mainly relates to image super-resolution reconstruction, in particular to an image super-resolution reconstruction method based on representation learning and neighborhood constraint embedding, which can be used for super-resolution reconstruction of various natural images . Background technique [0002] Image super-resolution reconstruction is to reconstruct a higher-resolution image from one or more low-resolution images. This method overcomes the limitations of resolution and high cost of traditional image sensors, and is a low-cost and efficient method to improve imaging quality, so it is very important in the fields of video, imaging, remote sensing, medicine, monitoring and military. Applications. Traditional super-resolution reconstruction methods include iterative back-projection method, maximum a posteriori probability method (MAP), maximum likelihood estimation method, co...

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): G06T5/50G06T3/40
Inventor 杨淑媛焦李成张继仁刘红英熊涛马晶晶缑水平刘芳侯彪刘正康崔顺
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