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

Feature layer fusion method and device based on graph embedding canonical correlation analysis

A typical correlation analysis and fusion method technology, applied in the field of feature layer fusion, can solve problems such as poor recognition effect, achieve good recognition effect, improve feature discrimination, and feature fusion effect

Pending Publication Date: 2020-06-26
ANHUI UNIVERSITY
View PDF4 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] In order to solve the technical problem of poor recognition effect of the existing feature fusion method, the present invention provides a feature layer fusion method and its device based on graph embedding canonical correlation analysis

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
  • Feature layer fusion method and device based on graph embedding canonical correlation analysis
  • Feature layer fusion method and device based on graph embedding canonical correlation analysis
  • Feature layer fusion method and device based on graph embedding canonical correlation analysis

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0065] see figure 1 , this embodiment provides a feature layer fusion method based on graph embedding canonical correlation analysis, which can be applied in speech recognition, emotion recognition, medical image analysis, and multimedia event detection. In this embodiment, the feature layer fusion method is mainly divided into two stages. The first stage selects representative representations of multiple single-mode features, and the second stage establishes correlations between multi-mode features.

[0066] In this embodiment, it is assumed that the j-th sample of the i-th modality that needs to be fused is n is the number of samples per modality, i.e. the sample size. P represents the number of heterogeneous sample sets that need to be fused, di represents the i-th modality sample dimension, and c is the number of categories. for any vector For a matrix A∈R d×n , the i-th row of A is denoted as a i , column j is denoted as a j , so Wherein, the feature layer fus...

Embodiment 2

[0122] This embodiment provides a feature layer fusion method based on graph embedding canonical correlation analysis, and a simulation experiment is performed on the basis of Embodiment 1. In the experimental part of this embodiment, several sets of experiments are proposed to verify the effectiveness of the feature layer fusion method. These groups of experiments are combining multiple features extracted in a single modality and combining features extracted separately in different modalities. Among them, the experiment is divided into three parts. Parts A and B combine the experimental results of different feature vectors extracted by a single mode, respectively using the multi-feature handwriting data set in the UCI data set and the Hong Kong Polytechnic University finger vein database. Part C is the experimental results of combining different biological characteristics of the hand, using the finger vein database provided by the USM database, the palmprint public database a...

Embodiment 3

[0190] This embodiment provides a feature layer fusion device based on graph embedding typical correlation analysis, which applies the feature layer fusion method based on graph embedding typical correlation analysis of embodiment 1 or embodiment 2, and includes a feature selection module, a similar graph matrix Building blocks as well as fusion modules.

[0191] The feature selection module is used to map the samples in all modalities to the projection matrix of the space of the same classification result, and apply L21 norm regularization to the projection matrix to select independent complementary features in multiple unimodal feature spaces at the same time Among them, the feature selection module includes an initial objective function establishment unit, an optimization unit and an update unit. The initial objective function establishing unit is used to establish the initial objective function. The initial objective function is used to learn the projection matrices for v...

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 feature layer fusion method and device based on graph embedding canonical correlation analysis. The feature layer fusion method comprises the steps that samples in all modesare mapped to a projection matrix of a space of the same classification result, and L21 norm regularization is applied to the projection matrix so that independent complementary features can be selected from multiple single-mode feature spaces at the same time; constructing a data similarity graph matrix to represent the similarity relation of sample points in the single-mode feature space; and learning a corresponding projection matrix for each mode through the regularization target function, and projecting the plurality of mode data to a projection subspace with the maximum discrimination, the maximum correlation and the minimum redundancy to realize multi-mode data fusion. According to the method, multi-modal data fusion is realized, the interference of redundant information in an original feature space is eliminated, the single-modal feature discriminability is improved, the correlation between multi-modal sample sets is enhanced, the recognition performance and stability are improved, the feature fusion effect is good, and the recognition effect is good.

Description

technical field [0001] The present invention relates to a feature layer fusion method in the technical field of data feature fusion, in particular to a feature layer fusion method based on graph-embedded canonical correlation analysis, and to a feature layer fusion device based on graph-embedded canonical correlation analysis using the method. Background technique [0002] Biometric recognition technology has the characteristics of uniqueness and stability, and has been applied to various authentication scenarios, such as finger vein recognition, iris recognition, gesture recognition, face recognition, etc. Although these identification methods have achieved good results, it is difficult to meet the application requirements in high-security fields because only a single biometric feature is used. It is very important to fuse and analyze multiple modal data. Multimodal fusion has been well applied in many fields, including audio-visual speech recognition, emotion recognition, ...

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
IPC IPC(8): G06K9/62G06K9/00
CPCG06V40/70G06F18/21322G06F18/21328G06F18/253Y02T10/40
Inventor 余程年王华彬申燕兰江浩李鑫王雨情施余峰陶亮
Owner ANHUI 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