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

Multi-object segmentation and tracking method based on dense track voting

A trajectory and dense technology, applied in the field of multi-target tracking, can solve the problem of low tracking accuracy and achieve the effect of simple and accurate segmentation and tracking

Inactive Publication Date: 2017-06-13
RES & DEV INST OF NORTHWESTERN POLYTECHNICAL UNIV IN SHENZHEN +1
View PDF4 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] In order to overcome the deficiency of low tracking accuracy of existing multi-target tracking methods, the present invention provides a multi-target segmentation and tracking method based on dense trajectory voting

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
  • Multi-object segmentation and tracking method based on dense track voting
  • Multi-object segmentation and tracking method based on dense track voting
  • Multi-object segmentation and tracking method based on dense track voting

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0039] The specific steps of the multi-target segmentation and tracking method based on dense trajectory voting in the present invention are as follows:

[0040] 1. Foreground extraction and target connected domain extraction.

[0041] Use the Vibe method to frame the current video frame j Extract the foreground and generate the foreground binary image; then use the connected domain processing of polynomial fitting to get the connected area Connect of the target i , and connect the connected domain Connect i Carry out number L i .

[0042] 2. Dense trajectory marking and cumulative voting.

[0043] After each connected domain number, according to the current jth frame Frame j Among them, the i-th dense trajectory Trajectory i endpoint v i Connected area Connect i marked with the corresponding designation T i = L i , an endpoint label T that does not belong to any connected domain i=-1, i is the track count number. Each node on the trajectory carries the label infor...

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 multi-object segmentation and tracking method based on dense track voting, which is used for solving the technical problem of low tracking accuracy rate of the existing multi-object tracking method. The technical scheme comprises the following steps of firstly, utilizing a foreground extraction method to extract a target connected domain, and then marking each track in each frame according to the connected domain where the track is located; and computing correlation between the tracks according to history marking information of the tracks, and then utilizing a region growing graph cut method to extract moving objects and carrying out division and tracking. According to the method, a tracking problem of the multiple moving objects is transformed to a weak classifier accumulation problem, and the division and the tracking of the multiple moving objects are simpler and more precise.

Description

technical field [0001] The invention relates to a multi-target tracking method, in particular to a multi-target segmentation and tracking method based on dense trajectory voting. Background technique [0002] Multi-target tracking technology is a traditional problem in the field of computer vision, and it is also a difficult problem. The document "Robust object tracking by hierarchical association of detection responses. In European Conference on Computer Vision, pages 788–801." discloses a multi-object tracking method. This method proposes a basic framework of low-to-high multi-layer tracking, the most critical of which is data association. The current methods based on data association have encountered unprecedented challenges, especially the occlusion problem. When objects in the scene are occluded, the detector cannot detect the target, and the trajectory will be interrupted. Most of the existing solutions are mostly based on the assumption of uniform motion of the tar...

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): G06T7/215G06T7/246G06T7/269G06T7/11
CPCG06T2207/10016
Inventor 杨涛张卓越张艳宁李治刘小飞
Owner RES & DEV INST OF NORTHWESTERN POLYTECHNICAL UNIV IN SHENZHEN
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com