Image matching method adopting window-equipped dynamic space structuring

A graph matching and spatial technology, applied in the field of computer vision and pattern recognition, can solve problems such as difficulty in obtaining the optimal objective function, and achieve the effect of low memory consumption and high matching accuracy

A graph matching and spatial technology, applied in the field of computer vision and pattern recognition, can solve problems such as difficulty in obtaining the optimal objective function, and achieve the effect of low memory consumption and high matching accuracy

CN106023232AInactive Publication Date: 2016-10-12UNIV OF ELECTRONICS SCI & TECH OF CHINA
6 Cites 1 Cited by

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Image matching method adopting window-equipped dynamic space structuring
  • Image matching method adopting window-equipped dynamic space structuring
  • Image matching method adopting window-equipped dynamic space structuring

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0031] figure 1 It is a flowchart of the graph matching method with windowed dynamic space regularization of the present invention.

[0032] Select the face picture in the Willow-Object database for experiment in the present embodiment, as figure 1 As shown, the present invention is based on the graph matching method of band window dynamic space regularization, comprises the following steps:

[0033] 1. Extract the image features of the reference picture image and the image to be matched respectively, and the features may include: single point features, edge features or block features, etc. Create a reference graph G 1 and the graph G to be matched 2 , where G 1 have N 1 nodes G 2 have N 2 nodes Establish reference graph G according to features 1 and the graph G to be matched 2 The ε-adjacency graph of points, such as figure 2 shown. Pre-set threshold ε, calculate the Euclidean distance d between every two feature points in the graph; when the distance is less ...

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 matching method adopting window-equipped dynamic space structuring and belongs to the field of machine vision and mode identification and specifically relates to the image matching method adopting window-equipped dynamic space structuring. According to local feature description of a reference image and an image to be matched, node to node matching degree can be measured via window-equipped dynamic space structuring, and therefore the image matching method is realized; the image matching method is characterized by low memory consumption and inexact matching. Based on an epsilon- adjacency map for a node and adjacent nodes thereof in the reference image G1, an image G2 is determined according to the number of the nodes, and sub-images to be matched are established; local information of a center node is expressed via included angle cosine in the sub-images, geometric matching between sub-images can be measured via window-equipped dynamic space structuring, center node matching and supporting degree can be obtained, a node pair with a maximum matching and supporting degree is selected, and then image matching processes are completed.

Description

technical field [0001] The invention belongs to the field of computer vision and pattern recognition, and more specifically relates to a graph matching method with window dynamic space regularization. Background technique [0002] Finding the mapping relationship between points and points from two point sets has always been the core issue in the field of computer vision and pattern recognition, that is, feature matching, and its applications are found in target recognition, retrieval, positioning, 3D reconstruction, motion segmentation, image deformation, etc. . In recent years, the research on feature matching using graph matching has developed rapidly. The graph can be constructed by extracting two-dimensional and three-dimensional feature points of the target, and then adding weighted edges. This is mainly because: graph matching methods can not only consider the first-order appearance information of features, but also use higher-order, such as side information, for fea...

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
12 Oct 2016
Publication
CN106023232A
IPC
G06T7/00
Inventors
郑亚莉; 程洪