Pedestrian recognition method based on cross modal comparison between image and video

A pedestrian re-identification, cross-modal technology, applied in the field of computer vision and pattern recognition, can solve problems such as reducing the recognition accuracy

Active Publication Date: 2017-12-15
暗物智能科技(广州)有限公司
View PDF4 Cites 22 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] However, how to effectively extract and reasonably utilize video information in the problem of pedestrian re-identification based on image and video comparison is one of the difficulties
Because compared to images, there is a lot of redundant information in videos, if not handled properly, it will reduce the accuracy of recognition
In addition, since the comparison between images and videos belongs to two different modalities, how to reasonably perform cross-modal comparison is another difficulty

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 recognition method based on cross modal comparison between image and video
  • Pedestrian recognition method based on cross modal comparison between image and video
  • Pedestrian recognition method based on cross modal comparison between image and video

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0038] The technical solutions of the present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0039] The present invention provides a pedestrian re-identification method based on cross-modal comparison between images and videos, which is used to retrieve videos containing corresponding persons in input query images from multiple videos. It should be noted that the multiple videos mentioned in the present invention may be multiple videos that are uniformly stored in a video database, or multiple videos that are distributed and stored.

[0040] Such as figure 1 As shown, a pedestrian re-identification method based on image and video cross-modal comparison provided by the present invention includes the following steps:

[0041] S1. Build a configurable depth model;

[0042] The depth model includes a convolutional neural network, a long-short-term memory network and a similarity learning network; the convolut...

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 provides a pedestrian recognition method based on the cross modal comparison between image and video, and is used for retrieving a video containing the corresponding characters in an input query image from multiple videos. The method includes the steps of S1, building a configurable depth model; S2, acquiring training samples, inputting the training samples into the depth model, training the depth model, and learning various parts of the parameters of the built depth model by utilizing the forward algorithm and the backward algorithm; S3, initializing the depth model by utilizing the obtained parameters learned in S2; inputting the query image and multiple videos to be measured in the depth model, and calculating the similarity measure between each video and the query image by utilizing the depth model; S4, listing the video with one threshold value higher than the similarity measure of the query image, and sorting according to the size of the similarity measure. According to the pedestrian recognition method, the pedestrian recognition based on the cross modal comparison between image and video under the precondition of guaranteeing high precision is achieved.

Description

technical field [0001] The invention relates to the fields of computer vision and pattern recognition, in particular to a pedestrian re-identification method based on cross-modal comparison of images and videos. Background technique [0002] Pedestrian re-identification technology is an important basic research topic in the field of computer vision. Person re-identification originated from the person tracking technology in the video field. When the tracked person temporarily leaves the camera shooting area, when he re-enters the shooting area, he needs to be re-identified and assigned the same ID as before. With the wide application of video surveillance, research on person re-identification has received more and more attention. At present, pedestrian re-identification is not limited to the recognition of the same person from a single perspective, but more generally refers to the re-identification of people at different times and from different perspectives. [0003] Most...

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): G06F17/30G06K9/00G06K9/62G06N3/04
CPCG06F16/738G06F16/784G06V40/103G06N3/045G06F18/22G06F18/214
Inventor 林倞张冬雨吴文熙
Owner 暗物智能科技(广州)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
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