A Metric Learning Method for Person Re-ID

A person re-recognition and metric learning technology, applied in character and pattern recognition, instrument, computing and other directions, can solve the problem of over-fitting of metric learning technology, and achieve the effect of strong generalization ability, excellent measurement effect, and excellent effect performance.

Inactive Publication Date: 2019-06-28
HUAZHONG UNIV OF SCI & TECH
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The present invention proposes a metric learning method for pedestrian re-identification, with the purpose of providing a metric learning method based on isometric constraints and solving the over-fitting problem existing in existing metric learning techniques

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 Metric Learning Method for Person Re-ID
  • A Metric Learning Method for Person Re-ID
  • A Metric Learning Method for Person Re-ID

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0028] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the following further describes the present invention in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0029] The following first explains and describes the terms used in the present invention.

[0030] Positive semi-definite matrix: an n*n Hermitian matrix M is positive definite if and only if for every non-zero complex vector z, there is z * Mz> 0, then M is called a positive definite matrix, where z * Represents the conjugate transpose of z. When z * Mz> 0 is reduced to z * When Mz≥0, M is said to be...

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 metric learning method for pedestrian re-identification, which comprises the following steps: firstly, a set of pedestrian target features under different cameras is established; The feature metric distance constraint is 0, and the feature distance constraint of different targets is a constant μ (μ>0); finally, the loss function is established and its structure is optimized, and the optimal metric matrix that satisfies the constraint conditions is obtained through the projection gradient descent method iteratively. Measuring the learning process. The invention effectively solves the problems of overfitting phenomenon and the sensitivity of the measurement matrix to noise in the existing metric learning method, and is suitable for the application occasion of pedestrian re-identification in complex scenes.

Description

Technical field [0001] The present invention belongs to the technical field of pattern recognition, and more specifically, relates to a metric learning method and system for pedestrian re-identification. Background technique [0002] Pedestrian re-recognition algorithm is one of the important areas of image processing and pattern recognition research, focusing on the recognition of specific target pedestrians under cameras without public view. At present, the more common method is to find a feature expression based on the appearance of pedestrians, which mainly includes information such as color and texture, and then use an appropriate measurement method to calculate the similarity between targets, and sort the output. Because the same target is affected by factors such as angle of view, illumination, and object occlusion under different cameras, its feature expression under different viewing angles is often biased. In addition to selecting a more robust feature expression target...

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
CPCG06V40/103G06F18/2132
Inventor 桑农陈科舟王金高常鑫李志强李亚成
Owner HUAZHONG UNIV OF SCI & TECH
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