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

A Method of Computing Image Local Feature Descriptor

A technology of feature descriptors and local features, applied in the sub-field of calculating local feature descriptors of images, can solve the problems of not considering the relevance of the angle, only considering the relevance, etc., to achieve a simple concept, strong discriminative ability and robustness. Effect

Active Publication Date: 2019-10-29
BEIJING UNION UNIVERSITY
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This invention only considers the relevance of the position, and does not consider the relevance of the angle

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
  • A Method of Computing Image Local Feature Descriptor
  • A Method of Computing Image Local Feature Descriptor
  • A Method of Computing Image Local Feature Descriptor

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0042] Such as figure 1 As shown, first calculate the position correlation matrix G of the feature descriptor. Execute step 100, initialize the position correlation matrix G=zeros(N 2 ,N 2 ), in this embodiment, set N=4, then G=zeros(16,16), that is, G is a 16×16 matrix, initialized with all elements being 0. Execute step 110, set parameter matrix G1 and G2, wherein Execute step 120, initialize setting matrix X=zeros(N 2 ,N 2 ), in this embodiment, N=4 is set, then X=zeros(16,16), that is, X is a 16×16 matrix, and all elements are initialized to be 0. Step 130 is executed to calculate the matrix X according to the parameter matrix. for i=1:16, for j=1:16, tmp=[G1(i)G2(i)]-[G1(j)G2(j)], X(i,j)=tmp×tmp T ,. The meaning of this formula is that when i is 1, 2, ..., 16 respectively, take j as 1, 2, ..., 16, calculate the tmp value respectively, and calculate the values ​​of all elements of the matrix X. Step 140 is executed to calculate the position correlation matrix G acc...

Embodiment 2

[0045] A 128-dimensional sift feature is:

[0046]

Embodiment 3

[0048] When σ=0.6, calculate the position correlation matrix G, G is a 16x16 diagonal matrix, the result is as follows:

[0049]

[0050]

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 present invention provides a method for calculating local feature descriptors of an image, which further includes the following steps: calculating a position correlation matrix G of the feature descriptors; and calculating an angle correlation matrix B. The present invention proposes a method for calculating local feature descriptors of images, and changes the SIFT descriptors through the transformation matrix, so that the original feature descriptors have more spatial information, so that the new descriptors have stronger discrimination ability and robustness.

Description

technical field [0001] The invention relates to the technical field of image feature description, in particular to a method for calculating image local feature descriptors. Background technique [0002] Image features are descriptions of image characteristics or attributes. The extraction and presentation of image features is the basis of image processing. Feature extraction is very important when representing images. Currently, a very important part of the features is the local features of the image. The extraction of local features includes feature point detection and feature point description. At present, there are a large number of local positive extraction methods, but the criteria for judging what is a good feature depends on the application, and there is no unified standard to measure it. [0003] The main purpose of feature point detection is to determine the position and scale of feature points and other parameters. The output of feature detection is a collection o...

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): G06K9/46G06F17/16
CPCG06F17/16G06V10/462
Inventor 马楠
Owner BEIJING UNION UNIVERSITY
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