Unlock instant, AI-driven research and patent intelligence for your innovation.

Unsupervised pedestrian re-identification method based on spherical similarity hierarchical clustering

A pedestrian re-identification and hierarchical clustering technology, applied in the field of pedestrian re-identification, can solve the problems of costing a lot of labor costs, and achieve the effects of saving labor costs, improving accuracy, and reducing intervention

Active Publication Date: 2020-04-07
CHINA UNIV OF MINING & TECH
View PDF5 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] In the current field of person re-identification, due to the requirements for data diversity, it usually takes a lot of labor costs to label the data set

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
  • Unsupervised pedestrian re-identification method based on spherical similarity hierarchical clustering
  • Unsupervised pedestrian re-identification method based on spherical similarity hierarchical clustering
  • Unsupervised pedestrian re-identification method based on spherical similarity hierarchical clustering

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0057] The present invention will be further described below in conjunction with the accompanying drawings.

[0058] Such as figure 1 Shown is a block diagram of the implementation process of an unsupervised person re-identification method based on hierarchical clustering of spherical similarity. The following will be described in detail with reference to the accompanying drawings.

[0059] Step 1. Input pedestrian image data

[0060] At present, the commonly used large-scale pedestrian re-identification datasets include Market-1501, Duke-MTMC, Mars, etc., all of which contain a large number of pedestrian images and the pedestrian labels corresponding to each image, which can be used as test data to verify the effectiveness of the present invention .

[0061] Step 2. Train the deep neural network and evaluate the person re-identification results

[0062] (21) Use the deep residual network to extract the features of the picture to obtain the features of the pedestrian pictur...

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 an unsupervised pedestrian re-identification method based on spherical similarity hierarchical clustering, and the method comprises the steps: employing spherical similarity tomeasure the similarity between pedestrian pictures, and weakening the structured difference, caused by the different positions of cameras, of light and angles; and after the picture feature vectors are normalized and compressed to the spherical surface, distinguishing the classification of the features more clearly and clearly. According to the method, a hierarchical clustering method based on spherical similarity and a method for image classification through spherical feature compression are integrated, pedestrian images can be retrieved on the premise of not manually labeling the images, and the high retrieval accuracy is achieved.

Description

technical field [0001] The invention relates to an unsupervised pedestrian re-identification method based on hierarchical clustering of spherical similarity, which uses spherical similarity to measure the similarity between pictures of pedestrians, and adds the internal average square deviation of the image category as a correction item, through the spherical features of the picture Classifying pictures belongs to the pedestrian re-identification technology in the field of image processing. Background technique [0002] In surveillance video, due to background occlusion and low resolution caused by pedestrians far away from the camera, it is often impossible to obtain pictures that can be used for face recognition. When face recognition technology cannot be used normally, pedestrian re-recognition becomes a very important substitute technology. A very important feature of pedestrian re-identification is cross-camera, so when evaluating performance in academic papers, it is ...

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): G06F16/783G06F16/55G06K9/62G06N3/04G06N3/08G06F17/15
CPCG06F16/784G06F16/55G06N3/088G06F17/15G06N3/045G06F18/231G06F18/22Y02T10/40
Inventor 周勇郑沂赵佳琦夏士雄姚睿杜文亮王秋张曼
Owner CHINA UNIV OF MINING & TECH