Image matching calculation method based on regional Gaussian weighting

A technique for matching computation and images, applied in the field of pattern recognition

Active Publication Date: 2014-12-10
北京博研高科技术有限公司
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to solve the limitation problem existing in the existing image matching technology, a kind of image matching calculation method based on region Gaussian weighting

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 matching calculation method based on regional Gaussian weighting
  • Image matching calculation method based on regional Gaussian weighting
  • Image matching calculation method based on regional Gaussian weighting

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0017] Attached below figure 2 , with image 3 As shown in the two images, they belong to the same web page. Because a large part of the page is flash, the similarity between the two images decreases. However, people’s subjective feeling is that the two images still come from the same page, that is, they should belong to the same category. The calculated results are in line with human subjective feelings, and the influence of flash needs to be removed.

[0018] The specific calculation process is as follows:

[0019] The first step is to align the images of the same type after normalization, and find the flash area by difference. According to the image structure, we first do edge detection, then use hough transform to find the top straight line in the image, align the two images, eliminate the deviation caused by the screenshot, and finally calculate the position of the flash by making a difference between the two images.

[0020] The second step is to find out the dissimi...

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 provides an image matching calculation method based on regional Gaussian weighting. Discrimination weights of all sub-regions of an image are obtained through advanced training, LBP texture column diagrams and color column diagrams of each sub-block are counted, and then final weighting similarity is obtained by multiplying and accumulating the column diagrams and the weights. The method is specifically realized as follows: firstly, finding out dissimilar regions of same types of images by performing background subtraction on normalized images; secondly, classifying image regions of different types of images normalized images into type similar regions, type dissimilar regions and middle regions according to distances of image sub-blocks and threshold values; thirdly, calculating the weights of the image sub-blocks through a Gaussian function; and fourthly, performing weighted accumulation on the texture column diagrams and the color column diagrams in the region according to solved regional weights, and calculating the similarity of the overall image.

Description

technical field [0001] The invention relates to a pattern recognition feature extraction method, which belongs to the technical field of pattern recognition. Background technique [0002] Image matching technology has a wide range of applications in many fields, such as image retrieval, phishing website identification and so on. Traditional image matching techniques have many limitations when matching similar webpage pictures, mainly because webpage images usually have certain similarities in structure and texture. When using traditional low-level features such as texture and color for matching, two images that are dissimilar to the human eye can easily be recognized by a computer as an image with a high degree of similarity, and the same web page image is sometimes affected by the page flash But it is easy to be judged as dissimilar images. However, high-level semantic features are not easy to obtain in existing technologies, so high-level features cannot solve the proble...

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 Patents(China)
IPC IPC(8): G06T7/00
Inventor 薄占滨
Owner 北京博研高科技术有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap