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

Hierarchical compressed image matching method and system based on orthogonal attention mechanism

A matching method and attention technology, applied in the field of graph computing and knowledge graph, can solve problems such as high time complexity and unbearable time

Active Publication Date: 2020-10-16
NAT UNIV OF DEFENSE TECH
View PDF4 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

But its time complexity is still very high, usually polynomial time or sub-exponential time of the number of nodes in the graph, and the time consumed by large graph matching is often unbearable

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
  • Hierarchical compressed image matching method and system based on orthogonal attention mechanism
  • Hierarchical compressed image matching method and system based on orthogonal attention mechanism
  • Hierarchical compressed image matching method and system based on orthogonal attention mechanism

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0081] Figure 1 to Figure 5 Shows a specific embodiment of the present invention, a hierarchical compression image matching method based on an orthogonal attention mechanism, which is mainly aimed at the matching problem of large images, such as figure 1 As shown, the logistic regression model used in this embodiment according to the scale of the processed data graph includes the following steps:

[0082] Step 1: Obtain the big picture data pair to be matched, and preprocess the big picture data. The preprocessing refers to the point vector initialization of the graph. The big picture data refers to the graph with more than 16 points; In this embodiment, a large picture from 20 to 500 points is used;

[0083] Step 2: Train the large-picture matching model based on the orthogonal attention mechanism according to the historical gallery; the specific training method of the large-picture matching model is:

[0084] Step 2.1: Obtain all the big picture data in the historical gallery, an...

Embodiment 2

[0117] The difference from the first embodiment is that this embodiment is mainly for the matching problem of small graphs, that is, for graphs with nodes less than 16 points, for small graphs, the prior art uses A* algorithm for graph matching However, this embodiment uses two layers of orthogonal attention and graph attention compression, and uses the compressed graph for graph matching.

[0118] A hierarchical compression graph matching method based on orthogonal attention mechanism, such as Image 6 As shown, including the following steps:

[0119] S1: Obtain the three-tuple graph data to be matched, and perform preprocessing on the small graph data. The preprocessing refers to the point vector initialization of the graph, and the small graph data refers to a graph whose number of nodes is less than 16 points;

[0120] S2: Train a small image matching model based on the orthogonal attention mechanism according to the historical gallery;

[0121] According to the method of claim 5,...

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 a hierarchical compressed graph matching method and system based on an orthogonal attention mechanism. The method comprises the following steps: acquiring a to-be-matched big picture data pair, and carrying out preprocessing on big picture data; training a big picture matching model based on an orthogonal attention mechanism according to a historical graph library; and inputting the preprocessed big picture data pair into the big picture matching model to obtain a matching result, and outputting the matching result. According to the invention, the picture attention network is used to carry out dimension reduction training on pictures in the process of acquiring picture vectors, point vectors are updated so as to allow the point vectors to better express the topological structures and node information of the pictures; the point vectors having undergone dimension reduction and an adjacent matrix are input into an orthogonal attention network for picture scale compression; through layer-by-layer compression, the extraction of picture information is finer, a more accurate picture vector is finally obtained; and then picture matching is performed through the compressed accurate picture vector, so the accuracy of a picture matching result is improved, a calculation amount is small, and calculation is quicker and more accurate.

Description

Technical field [0001] The invention belongs to the technical fields of graph computing and knowledge graphs, and particularly relates to a hierarchical compression graph matching method and system based on an orthogonal attention mechanism. Background technique [0002] Graph matching has a wide range of applications, such as protein structure matching, three-dimensional object matching, road network analysis and social network learning. Its essence is the problem of graph isomorphism, and graph isomorphism has been proved to be NP-complete, and there is no solution in polynomial time. Therefore, it is difficult to determine whether two graphs match. [0003] Generally, the research on graph matching in academia is divided into two categories: exact matching and approximate matching. People initially conducted in-depth research on exact matching, and produced a series of graph exact matching algorithms, among which the most representative algorithm is the A* algorithm. However, ...

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): G06K9/62G06N3/04
CPCG06N3/04G06F18/22
Inventor 李东升刘苧蹇松雷赖志权刘锋陈易欣黄震
Owner NAT UNIV OF DEFENSE TECH
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