Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Pedestrian re-identification method based on multi-layer fusion and alignment division

A pedestrian re-identification, multi-layer fusion technology, applied in the field of pedestrian re-identification, can solve the problem of ignoring the practical value of shallow features, and achieve the effect of improving performance

Pending Publication Date: 2020-11-03
上海蠡图信息科技有限公司
View PDF0 Cites 9 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Most of the current deep learning-based methods only use deep features, while ignoring the practical value of shallow features.

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 based on multi-layer fusion and alignment division
  • Pedestrian re-identification method based on multi-layer fusion and alignment division
  • Pedestrian re-identification method based on multi-layer fusion and alignment division

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0040] Pedestrian re-identification has a great demand in the field of intelligent security. It aims to associate the same pedestrians in different places at different times. The general approach is to give a picture of a pedestrian to be retrieved, extract the features of the query picture and the pictures in the gallery through the trained model, and sort the pictures in the gallery according to the similarity of feature embedding, so as to perform pedestrian image retrieval . In recent years, the task of pedestrian re-identification has made great progress. However, in an open outdoor environment, pedestrian images will have large differences due to the presence of interference such as pose, occlusion, clothing, background clutter, and camera perspective. Therefore, pedestrian re-identification Recognition is still a very challenging task. There are generally two directions to solve the problem of pedestrian re-identification, called representation learning and metric lear...

Embodiment 2

[0084] In order to verify the actual effect of the present invention, experiments are carried out on the three most commonly used datasets for pedestrian re-identification tasks: Marker-1501, DukeMTMC-reID, and CUHK03. The first successful matching probability (rank-1) and mean average precision (mAP) are used to evaluate the experimental results. Since the rank-1 standard will reach saturation when the matching reaches a certain level, the degree of discrimination is not high, so more emphasis is placed on the evaluation of mAP.

[0085] It should be noted that Marker-1501 includes 1501 pedestrians with different identities captured by 6 cameras, and the data set generates 32668 pictures containing individual pedestrians through the DPM detector. They are divided into non-overlapping training / test sets. The training set contains 12,936 images of 751 pedestrians with different identities. The test set contains 3,368 query images and 19,732 gallery images from 750 different ped...

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 based on multi-layer fusion and alignment division, and the method comprises the following steps: constructing a pedestrian re-identification network model, and training the pedestrian re-identification network model; fusing the feature maps of different layers with the feature map of the last layer by using a multi-layer fusion modulein the network model to obtain a multi-layer fusion feature finally containing shallow feature information; extracting the central position of the pedestrian by using an alignment division module in the network model, and expanding the central position to two sides to obtain the local features of the pedestrian accurately segmenting the local area; and connecting the multi-layer fusion features, the local features and the global features according to channel dimensions to obtain final pedestrian discrimination features to complete pedestrian re-identification. The invention has the beneficialeffects that the proposed fusion module can fuse information carried by feature maps of different levels, and on the basis, multi-layer fusion features are extracted and added into final discrimination features for auxiliary identification, so the re-identification performance is effectively improved.

Description

technical field [0001] The invention relates to the technical field of pedestrian re-identification, in particular to a pedestrian re-identification method based on multi-layer fusion and alignment division. Background technique [0002] In recent years, deep learning methods have been widely used in pedestrian re-identification tasks. The mainstream method is to extract features containing high-level semantic information through deep networks for recognition. However, simply using the high-level semantic information (objects or parts) contained in deep features often ignores the information (color, texture, etc.) Prominent pedestrians can be identified even when the image resolution is very low, and shallow information is very effective for identifying such images. For images with less obvious features, it is necessary to use deep networks to extract high-level semantics (some prominent images of the body). Most of the current deep learning-based methods only use deep fea...

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
CPCG06N3/08G06V40/103G06N3/045G06F18/213G06F18/253
Inventor 宋晓宁王鹏冯振华
Owner 上海蠡图信息科技有限公司
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
Eureka Blog
Learn More
PatSnap group products