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

A transfer learning method for single-sample face recognition based on lpp feature extraction

A technology of transfer learning and feature extraction, applied in the field of pattern recognition

Active Publication Date: 2016-12-07
南京数字空间新技术研究院有限公司
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The improvement of generalization ability can improve the accuracy of single-sample recognition to a certain extent, but it cannot fundamentally solve the contradiction between small sample number and high dimensionality

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
  • A transfer learning method for single-sample face recognition based on lpp feature extraction
  • A transfer learning method for single-sample face recognition based on lpp feature extraction
  • A transfer learning method for single-sample face recognition based on lpp feature extraction

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0055] Embodiments of the present invention will be described in detail below in conjunction with specific drawings and examples.

[0056] Such as figure 1 As shown, a single-sample face recognition transfer learning method based on LPP feature extraction includes the following steps:

[0057] Step 1, given the migration source TS, calculate the average face AF of category i i , and based on the prior probability to solve the intra-class sample covariance matrix ∑ w , and obtain the whitening operator W w ;

[0058] Intra-class sample covariance matrix ∑ based on prior probability w The expression is as follows:

[0059] Σ w = Σ i = 1 L P ( I i ) Σ s = 1 K ...

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 belongs to the technical field of pattern recognition, and in particular relates to a single sample face identification transfer learning method based on LPP feature extraction. The method is different from a traditional global face recognition method based on generalization ability improvement and a traditional local face recognition method depending on image segmentation. The method provided by the invention comprises the steps that whitening cosine similarity is used to screen a migration source to acquire a selected sample source; feature projection is respectively carried out on source feature and target feature faces in the selected source by using LPP, a feature migration matrix is solved, and mapping relationship approximating is carried out; the feature migration matrix is imposed on a training sample, and the original macro feature is migrated to a goal macro feature; and face recognition with high accuracy is realized by using a nearest classifier. According to the invention, a number of source samples correlated with a target single sample are effectively used; rational screening and macro feature migration are carried out; the problem of difficult single sample training is solved to a large extent; and high face recognition accuracy can be acquired.

Description

technical field [0001] The invention belongs to the field of pattern recognition, and in particular relates to a single-sample face recognition method based on LPP (Locality Preserving Projections, Locality Preserving Projections) feature migration. Background technique [0002] As a typical high-dimensional small-sample problem, face recognition has important applications in smart card design, access control, information security, and law enforcement tracking. There is only one training sample for a face, and the test sample is affected by factors such as expression, lighting, and angle, and often has a large difference from the training sample. This has led to certain difficulties in the further promotion and application of face recognition technology, and conventional transfer learning methods are difficult to deal with this problem. Technically speaking, single training sample face recognition refers to identifying the identity of a person in an image whose pose, lighti...

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): G06K9/00G06K9/62
Inventor 王雪松潘杰程玉虎
Owner 南京数字空间新技术研究院有限公司
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