Pedestrian re-recognition method based on transfer learning

A pedestrian re-identification and transfer learning technology, applied in the field of intelligent monitoring, can solve the problems of low accuracy, low accuracy of pedestrian re-identification, and few samples.

Active Publication Date: 2015-11-25
CHINA JILIANG UNIV
View PDF4 Cites 27 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] Aiming at a series of problems in the prior art, such as the low accuracy rate and low precision of pedestrian re-identification in the monitoring network, and when the target pedestrian appears infrequently under the known camera, there are relatively fe

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-recognition method based on transfer learning
  • Pedestrian re-recognition method based on transfer learning
  • Pedestrian re-recognition method based on transfer learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0030] The present invention will be further described below in conjunction with accompanying drawing. Such as figure 1 , 2 , As shown in 3, the pedestrian re-identification method based on transfer learning of the present invention comprises steps as follows:

[0031] Step 1. Pedestrian foreground segmentation. The GrabCut algorithm is used to perform known pedestrian foreground segmentation on each frame of all camera video sequences with known pedestrians. This algorithm uses the texture (color) information and boundary (contrast) information in the image, and then specifies some pixels to belong to target, a better segmentation result can be obtained. Also, we already know the camera presence of these known pedestrians.

[0032] Step 2, pedestrian feature extraction. Use the human body symmetry model to divide the extracted known pedestrians into five regions of interest: head, left upper limb, right upper limb, left leg, and right leg, and extract the color features,...

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-recognition method based on transfer learning and aims at solving a problem that pedestrian samples to be re-recognized are less. The method has the following steps of (1) pedestrian prospect segmentation: acquiring known pedestrians in different cameras under different scenes; (2) feature extraction: extracting features of the known pedestrians; (3) improved nerve network model learning in a source field: learning through the improved nerve network model and acquiring a model parameter which is shown in the description; (4) transfer learning from the source field to a target field: through transferring the model parameter to an improved nerve network model parameter in the target field, learning and acquiring a target field model parameter; (5) pedestrian re-recognizing: using the improved nerve network model in the target field to carry out pedestrian re-recognition discrimination.

Description

technical field [0001] The invention belongs to the field of intelligent monitoring, and in particular relates to a pedestrian re-identification method in monitoring video. Background technique [0002] In recent years, video surveillance systems have become popular, especially in public security, and play a pivotal role in combating crime and maintaining social stability. Re-identification of specific targets in surveillance systems is an important part of the public security field. With the development of technology and the increase of application requirements, the re-identification problem is gradually developing into a research hotspot. However, under actual conditions, if the target to be re-recognized appears less frequently under the camera, the sample data that can be used is relatively small, which affects the establishment of the re-recognition model, and finally easily leads to wrong recognition of the target. Contents of the invention [0003] Aiming at a ser...

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
IPC IPC(8): G06K9/00
CPCG06V40/20G06V20/53
Inventor 章东平徐佳慧孙敏徐丽园
Owner CHINA JILIANG 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