Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Learning-Based Similarity Metrics

A similarity measurement and correlation technology, applied in the field of learning-based similarity measurement, can solve the problem that model-based methods cannot work

Active Publication Date: 2021-09-14
TIANJIN UNIV +1
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In this case, model-based methods cannot work and only image-based methods can be applied

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
  • Learning-Based Similarity Metrics
  • Learning-Based Similarity Metrics
  • Learning-Based Similarity Metrics

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0073] The most important part of the learning-based similarity measurement method is to extract two different relationship graphs of the model in the early stage, namely the view-based hypergraph and the model-based graph. For the view of the model, through HAC [11] (HAC is a kind of hierarchical clustering method, and its basic idea is: treat a single document as a different class, and then use different methods to merge it, so that the number of classes gradually decreases, until finally it is a class or clustered to all number of classes required) to build view clusters, and establish a view-based hypergraph according to the view clusters; for model data, use the spatial structure circular descriptor SSCD [12] (In SSCD, the spatial structure of the 3D model is described by a 2D image, and the attribute value of each pixel represents 3D spatial information, SSCD can preserve the global spatial structure of the 3D model with rotation and scaling invariance) The method extract...

Embodiment 2

[0094] Combined with the specific calculation formula, figure 1 , figure 2 The scheme in Example 1 is further introduced, see the following description for details:

[0095] Use O={O1 ,O 2 ,...,O n} represents n solid models, and V i ={v i1 ,v i2 ,...,v im} represents multiple views of the i-th stereoscopic model, from which a representative view is selected, assuming that the selected representative view is V i ={v i1 ,v i2 ,...,v im}, and then use the star expansion to construct the hypergraph of the stereo model, denoted as G H =(V H ,E H ,W H ); where V represents the vertex, E represents the edge, W is the weight of the edge E, and H is the incidence matrix.

[0096] Assume n stereo models have a total of n r representative views, first calculate the distance based on Zernike moment between every two representative views, and generate the top K nearest views for each representative view, the value of K is set to 10 in the embodiment of the present inventio...

Embodiment 3

[0127] Below in conjunction with specific example, the scheme in embodiment 1 and 2 is carried out feasibility verification, see the following description for details:

[0128] The database in the embodiment of the present invention is a database based on NTU and PSB [7] to carry out. These three-dimensional models are drawn by the staff through three-dimensional model processing software such as 3DMax or collected from websites with different domain names. The three-dimensional model database has different storage formats including *.obj and *.off formats. In this experimental design The representative NTU549 database is used, which contains a total of 549 three-dimensional models of 47 categories, the PSB database contains 161 categories, a total of 1814 three-dimensional models, and the SHREC database contains 40 categories, a total of more than 800 three-dimensional models. Some models used in the embodiment of the present invention are as attached figure 2 As shown, th...

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 similarity measurement method based on learning, which includes: given the views of the stereo model, screening representative views to construct a view-based hypergraph to represent the relationship between the stereo objects; using the stereo model data, Spatial structure circular descriptors are extracted from each stereotype, and the distance between each stereotype is used to generate a simple model-based graph to explore the correlation between stereotypes; an appropriate learning framework is selected, and an initial The learning weight of , the hypergraph generated based on the view and the graph generated based on the stereo model are used as the input of the learning framework, and the optimal combination weight of the two graphs is learned through the joint learning framework, so that through the hypergraph based on the view and the graph generated based on the stereo model graph to estimate the correlation between stereo objects. The invention makes the description of the three-dimensional model more comprehensive and more accurate and scientific in the aspect of similarity quantification by extracting the view feature information and the three-dimensional model space structure information.

Description

technical field [0001] The invention relates to the fields of similarity measurement, three-dimensional model retrieval and the like, in particular to a learning-based similarity measurement method. Background technique [0002] Due to the rapid development of graphics hardware, computer technology and network, three-dimensional objects have been widely used in various applications, such as: computer graphics, medical industry and virtual reality field. Large-scale databases of volumetric objects are increasing rapidly, leading to a gradual increase in the demand for efficient algorithms for volumetric object retrieval. [0003] Recently, extensive research work has been devoted to stereoscopic object retrieval techniques [1]-[4] . Existing stereoscopic object retrieval methods can be simply divided into two paradigms, model-based and view-based methods. [0004] In a model-based approach [5]-[7] In , volumetric objects are described as model-based features, e.g. low-lev...

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): G06K9/62
CPCG06F18/22G06F18/217
Inventor 王坤韦莎程雨航王伟忠聂为之刘安安苏育挺
Owner TIANJIN UNIV
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
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