Matrix index-based elasticity preserving projection method and application thereof

A matrix index and projection-preserving technology, which is applied to instruments, character and pattern recognition, computer components, etc., can solve the problems of sensitive neighborhood parameters and unconsidered data structure information, so as to improve robustness and reduce sensitivity Effect

Pending Publication Date: 2020-04-17
INFORMATION SCI RES INST OF CETC
View PDF3 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Aiming at the small sample problem (singular value problem), sensitivity to neighborhood parameters, and no consideration of the structural information of the data itself in the current dimensionality reduction algorithm, the present invention proposes an elastic-preserving projection method based on matrix index

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
  • Matrix index-based elasticity preserving projection method and application thereof
  • Matrix index-based elasticity preserving projection method and application thereof
  • Matrix index-based elasticity preserving projection method and application thereof

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0017] The following description serves to disclose the present invention to enable those skilled in the art to carry out the present invention. The preferred embodiments described below are only examples, and those skilled in the art can devise other obvious variations.

[0018] For the face recognition problem, given a face image training sample set X={x 1 , x 2 ,...,x N}∈R D×N , x i Represents a D-dimensional feature vector drawn from a face image. For example, for a 100- to 100-pixel face image, it will be a D=10000-dimensional feature vector after expansion. N represents the number of training samples. The purpose of the present invention is to find a transformation matrix W∈R D×d , and use this transformation matrix to get the low-dimensional embedding of face data: y i =W T x i ,y i ∈R d×N , where d

[0019] The e...

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 matrix index-based elasticity preserving projection method and application thereof, and the method comprises the following steps: in order to capture local and global structure characteristics of image data, respectively constructing local and global neighborhood graphs, and calculating weights to obtain similar matrixes Slocal and Sglobal; calculating a Laplace matrix according to the similarity matrix; performing matrix exponent operation and defining a target function; performing objective function optimization and solving operation. On one hand, essential structure information of data is mined by constructing local and global graphs; and on the other hand, matrix exponent calculation is introduced to enable a matrix to be in a full rank, so that the problem ofsmall samples is solved, and meanwhile, after the matrix exponent is introduced, a small characteristic value caused by neighborhood parameter change is weakened, and a relatively large characteristic value can be amplified, so that the robustness of the algorithm to domain parameters can be enhanced.

Description

technical field [0001] The invention discloses a dimension reduction method-exponential elasticity preserving projection, which belongs to the technical field of biological feature extraction and pattern recognition, and involves the construction of local and global neighborhood structures of data, matrix index calculation, and optimization of objective functions, and can be used for Image recognition, data mining, data clustering. Background technique [0002] Principal Component Analysis (PCA) and Linear Discriminant Analysis (Linear Discrimination Analysis, LDA), as typical dimensionality reduction methods, have been widely used in many fields. The main goal of PCA is to find the projection direction that maximizes the data covariance. The projection direction is a set of optimal unit orthogonal vector bases obtained through linear transformation. The linear combination of these vectors can reconstruct the original sample, and the reconstructed The error between the samp...

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): G06K9/00G06K9/62
CPCG06V40/172G06V40/168G06F18/22G06F18/213
Inventor 袁森葛建军黄文涛张峰
Owner INFORMATION SCI RES INST OF CETC
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
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap