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

Object identification method based on salient region bag-of-word model

A bag of words model, object recognition technology, applied in character and pattern recognition, instruments, computer parts, etc., can solve the problems of insufficient image feature expression and insufficient quantity.

Inactive Publication Date: 2017-04-05
BEIJING UNION UNIVERSITY
View PDF15 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the number of feature points detected by the general interest point detection operator has the defect of insufficient number, which makes the image feature expression insufficient, but the detected feature points are often concentrated in the target position.

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
  • Object identification method based on salient region bag-of-word model
  • Object identification method based on salient region bag-of-word model
  • Object identification method based on salient region bag-of-word model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0074] The object recognition method based on the salient region bag-of-words model includes three parts: salient region location, feature extraction and description, and image region feature similarity comparison.

[0075] Execute step 100 to perform corner detection.

[0076] ShiTomasi corner points are calculated by calculating the rate of change of the gradient direction, which is the point where the brightness of the image changes drastically or the curvature is very large. The main idea is to use the automatic correlation matrix to determine the change form of the signal in the image. Assuming that the signal at x in the image is I(x), using the image signal and the Gaussian function G(x, σ D ) to perform convolution operation to obtain the first derivative.

[0077] L(x,σ D )=I(x)*G u (x,σ D ) (1)

[0078] L v (x,σ D )=I(x)*G v (x,σ D ) (2)

[0079] L u (x,σ D ) L v (x,σ D ) = I u (x,σ D ) (3)

[0080] where σ D is the differential scale, using the for...

Embodiment 2

[0105] like figure 2 As shown, the similarity comparison of image region features includes the training and testing process of the images in the image library.

[0106] (1) Training process

[0107] Execute step 200, read in the training object picture, and determine the salient region of the image.

[0108] Execute step 210, extract the SIFT feature of the training sample in the salient area, if there are i pieces of training pictures, the number of SIFT feature points of each image is n 1， no 2 ,...,n i , the total number of extracted SIFT features is (n 1 +n 2 +…+n i ).

[0109] Execute step 211, use a size of (n 1 +n 2 +…+n i )×128. The original training matrix is ​​used to store the SIFT features of all samples, and the k-means clustering algorithm is used to create the visual dictionary required by the BOW model. k is the size of the visual dictionary, that is, the dimension of the BOW histogram. Execute step 230, map on the visual dictionary, count the BOW ...

Embodiment 3

[0135] like image 3 As shown, in order to find a stable key point in the multi-scale space, first construct each layer of the pyramid to build a Gaussian difference scale space (DoG):

[0136] D(x,y,σ)=(G(x,y,kσ)-G(x,y,σ))*I(x,y)=L(x,y,kσ)-L(x,y , σ)

[0137] The DoG difference pyramid is obtained from the Gaussian pyramid.

[0138] like Figure 4 As shown, the local extremum points detected on the DoG space are used as key points. In order to find the extremum points, each pixel needs to correspond to 8 adjacent points of the same scale as it and 9×2 adjacent points of the upper and lower adjacent scales. A total of 26 points are compared to ensure that an extreme point is detected in both the scale space and the two-dimensional image space. The pixel point is larger or smaller than the 26 points, that is, it is determined as a key point.

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 object identification method based on a salient region bag-of-word model; the method comprises the following steps: corner detection, locating an image salient region, SIFT characteristic extraction, and image region characteristic similarity comparison. The method can extract local characteristics in the target region, thus preventing application of complex image segmentation technology on one hand, and greatly reducing characteristic points having no relation with the object.

Description

technical field [0001] The invention relates to the technical field of digital image processing, in particular to an object recognition method based on a salient region bag-of-words model. Background technique [0002] With the rapid development of related technologies in the field of machine learning and pattern recognition, as well as the continuous improvement of computer vision technology, it has become possible to use computers to imitate human cognitive abilities and then reduce or assist people in completing daily work tasks. Object recognition has become an extremely important research direction in pattern recognition, and has a wide range of needs and applications in military and civilian fields. Such as: intelligent video surveillance, self-driving navigation, human-computer interaction, massive content-based image retrieval on the Internet, etc. [0003] How to accurately and effectively recognize objects to meet the increasing needs of machine vision, psychology...

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): G06K9/62
CPCG06F18/23213G06F18/2411
Inventor 袁家政刘宏哲郭燕飞
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