Supercharge Your Innovation With Domain-Expert AI Agents!

Template update target tracking algorithm based on multilayer features of fully convolutional twin network

A twin network and template update technology, applied in biological neural network models, computing, neural learning methods, etc., can solve the problems of template pollution, poor robustness of object surface deformation, etc., to improve accuracy, improve performance and tracking speed, Solve the effect of template pollution

Pending Publication Date: 2022-06-03
XIAN UNIV OF TECH
View PDF0 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a template update target tracking algorithm based on the multi-layer features of the fully convolutional twin network, which solves the problem of poor robustness to the deformation of the object's appearance during tracking in the prior art and the template is updated due to template update. pollution problem

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
  • Template update target tracking algorithm based on multilayer features of fully convolutional twin network
  • Template update target tracking algorithm based on multilayer features of fully convolutional twin network
  • Template update target tracking algorithm based on multilayer features of fully convolutional twin network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0084] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0085] The invention provides a template update target tracking algorithm based on multi-layer features of full convolution Siamese network, such as figure 1 shown, follow the steps below:

[0086] Step 1, construct the overall network, and perform end-to-end training on the overall network structure;

[0087] The overall network structure is divided into three parts: the first part is the Siamese neural network used for deep feature extraction, and the second part is the 3D convolutional neural network used for template update, that is, the 3D template update module. The first part and the second part are composed of Feature extraction network, the third part includes classification branch and regression branch;

[0088] The Siamese neural network is divided into four layers (P2, P3, P4, P5): the first two layers are composed of convolutiona...

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 template updating target tracking algorithm based on multilayer features of a fully convolutional twin network. The template updating target tracking algorithm specifically comprises the steps of constructing an overall network and performing training; performing initialization tracking setting on a video image sequence to be tracked by using the trained network to obtain initial position information of an initial target template and a target; entering a normal tracking process to obtain a tracking result response diagram of the current frame; a template updating condition judgment method based on standard mutual information is used for judging whether the current tracking result is reliable or not, if the current tracking result is reliable, a template is updated, if the current tracking result is not reliable, the template is not updated, and if the number of the reserved reliable tracking results is two, the latest result is used for replacing the oldest result; using the latest template to continuously perform normal tracking on the subsequent video image sequence of the currently tracked video frame; and repeating the steps 3-5 until all the video image sequences are tracked, thereby obtaining the position of the target in each frame of the video, and ending the tracking task.

Description

technical field [0001] The invention belongs to the technical field of video target tracking, and relates to a template update target tracking algorithm based on multi-layer features of a fully convolutional twin network. Background technique [0002] Object tracking is an important topic in the field of computer vision, which has far-reaching research significance and is widely used in intelligent video surveillance, unmanned driving, human-computer interaction and other fields. [0003] The single target tracking task refers to the process of locating the position of the target in subsequent frames according to the target tracking algorithm after given the target size and position information in the first frame of the video for a set of video image sequences. With the maturity of deep learning technology, researchers began to apply it to target tracking, and the target tracking algorithm based on deep learning based on twin neural network has gradually become a mainstream ...

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): G06T7/246G06N3/04G06N3/08
CPCG06T7/246G06N3/08G06T2207/10016G06T2207/20081G06T2207/20084G06N3/045
Inventor 鲁晓锋李小鹏王轩王正洋柏晓飞李思训姬文江黑新宏
Owner XIAN UNIV OF TECH
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