Pedestrian re-identification method and system based on mixed supervision

A pedestrian re-identification and pedestrian technology, applied in the field of computer vision, can solve difficult and unconsidered problems, and achieve the effect of reducing the impact of pseudo-label noise, reliable learning goals, and speeding up the feature learning process.

Pending Publication Date: 2021-11-12
SHANDONG NORMAL UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In recent years, pseudo-label methods have brought huge performance improvements for person re-identification; however, due to the limitations of their methods, labeled data may not be considered during training, and in order to ensure the reliability of pseudo-labels, the clustering process Outliers in are discarded and not used for training; however, these outliers may be difficult but valuable samples during training

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 and system based on mixed supervision
  • Pedestrian re-identification method and system based on mixed supervision
  • Pedestrian re-identification method and system based on mixed supervision

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0040] Such as figure 1 As shown, this embodiment provides a method for pedestrian re-identification based on hybrid supervision, including the following:

[0041] 1. Select and organize training images:

[0042] In this embodiment, two kinds of training images are used, one is a single-camera-labeled pedestrian image, and the other is an unlabeled pedestrian image; wherein, for the selection of a single-camera-labeled pedestrian image, specifically, randomly select a camera for each person , and the image under the selected camera is used as the training image; for the training sample X of the unlabeled pedestrian image t , using the self-step clustering strategy to divide the samples into clusters and cluster outliers; the entire training set is divided into, and the labeled samples under a single camera are X s , unlabeled cluster samples and cluster outlier samples three parts;

[0043] Preferably, during training, each mini-batch contains 64 single-camera annotat...

Embodiment 2

[0069] This embodiment provides a pedestrian re-identification system based on hybrid supervision, including: a sample building module, a clustering module, a training module and a dynamic updating module;

[0070] The sample building module is configured to: select a single camera to mark pedestrian images and unlabeled pedestrian images as training image samples;

[0071] The clustering module is configured to: cluster the unlabeled pedestrian images, segment them into clustered pedestrian images and non-clustered outliers, and judge the clustering reliability;

[0072] The training module is configured to: provide supervision signals of single-camera labeled pedestrian images, unlabeled clustered pedestrian images, and non-clustered outliers, and jointly train the supervisory signal and the encoder;

[0073] The dynamic update module is configured to: dynamically update single-camera annotated pedestrian images, unlabeled clustered pedestrian images, and supervision signals...

Embodiment 3

[0075] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored on the memory and operable on the processor. When the processor executes the program, the hybrid supervision based on Embodiment 1 is implemented. person re-identification method.

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 and system based on mixed supervision. The method comprises the following steps: selecting a single-camera labeled pedestrian image and an unlabeled pedestrian image as training image samples; clustering the unlabeled pedestrian image, segmenting the unlabeled pedestrian image into a clustered pedestrian image and a non-clustered abnormal value, and judging the clustering reliability; providing supervision signals of a single-camera labeled pedestrian image, an unlabeled clustered pedestrian image and a non-clustered abnormal value, and carrying out joint training on the supervision signals and an encoder; and dynamically updating supervision signals of the single-camera labeled pedestrian image, the unlabeled clustered pedestrian image and the non-clustered abnormal value. According to the invention, joint feature learning is carried out through all available information of the single-camera labeled data and the unlabeled data, and training is carried out by using the most reliable clustering of the unlabeled data, so that a more reliable learning target is provided.

Description

technical field [0001] The disclosure belongs to the technical field of computer vision, and in particular relates to a method and system for pedestrian re-identification based on hybrid supervision. Background technique [0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art. [0003] In surveillance video, due to camera resolution and shooting angle, it is usually impossible to obtain very high-quality face pictures; when face recognition fails, pedestrian re-identification becomes a very important substitute technology; Re-recognition is an important research direction after face recognition; the object of research is the entire characteristics of people, including clothing, body posture, hairstyle, posture, etc.; it can be used as a supplement to face recognition technology, and interact, applied to More scenes. [0004] The goal of person re-identification is to find th...

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
CPCG06F18/23213G06F18/214
Inventor 高玲张如梦李梦瑶
Owner SHANDONG NORMAL UNIV
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