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

Pedestrian re-identification method combined with picture generation

A pedestrian re-identification and image generation technology, applied in the field of computer vision, to achieve the effect of shortening intra-class differences, improving recognition accuracy, and increasing inter-class differences

Active Publication Date: 2021-07-16
BEIJING INSTITUTE OF TECHNOLOGYGY
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, in practical applications, pre-trained models with high recognition accuracy do not necessarily exist

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
  • Pedestrian re-identification method combined with picture generation
  • Pedestrian re-identification method combined with picture generation
  • Pedestrian re-identification method combined with picture generation

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0021] The present invention will be described in detail below with reference to the accompanying drawings and examples.

[0022] The present invention provides a pedestrian re-identification method generated by joint pictures, the flow chart of which is as follows figure 1 As shown, it specifically includes the following steps:

[0023] Step 1, read the data from the sample data set, and obtain the pedestrian image x in the data set i (x i ∈X, i=1,2...,N, N is the total number of pictures in the sample data set) and its label identity information y k (y k ∈Y, k=1,2...,K, K is the total number of identity categories in the sample data set).

[0024] Step 2, build a generative confrontation network (GAN):

[0025] Generative confrontation network (GAN) includes an encoder E, a generator G and a discriminator D, where the encoder includes a structural encoder E s and feature encoder E f ; The overall structure of the Generative Adversarial Network is as follows figure 2...

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 pedestrian re-identification method combined with picture generation. According to the method, pedestrian pictures are re-identified based on the GAN, and in the training process of the GAN, a similar unsupervised learning method is used, a teacher model is adopted to predict identities (feature information) of the pedestrian pictures and guide a feature encoder of the GAN, so that information of generated samples is fully utilized, the quality of the model is improved, the student model with higher precision is obtained. Extra annotation information, a pre-training model and the like do not need to be introduced, and the applicability of a pedestrian re-recognition algorithm in different scenes can be improved. In addition, the thought of triple loss is further utilized, data are expanded, intra-class differences are shortened, meanwhile, inter-class differences are increased, and the recognition precision is further improved.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a pedestrian re-identification method for joint image generation. Background technique [0002] Pedestrian re-identification refers to knowing a pedestrian picture to be retrieved, and retrieving other pictures of the pedestrian from a non-overlapping camera image database, which is usually regarded as a metric learning problem. The difficulty lies in that the pictures taken by different cameras have serious intra-class differences in pedestrian appearance, posture, background and other information. At the same time, pedestrians will also have certain inter-class similarities in a specific perspective. With the development of deep convolutional networks in recent years, more robust representations of pedestrian pictures and stronger recognition capabilities have made great progress in pedestrian re-identification, and even performed better than human recognition on some ...

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): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/088G06V40/103G06N3/045G06F18/22G06F18/2155Y02T10/40
Inventor 苏迪张成王少博邱语聃冀瑞静
Owner BEIJING INSTITUTE OF TECHNOLOGYGY
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