Supercharge Your Innovation With Domain-Expert AI Agents!

Block target tracking method on multi-domain convolutional neural network

A convolutional neural network and target tracking technology, which is applied in the field of block target tracking, can solve the problems of low deep learning performance and complexity, and achieve excellent tracking performance, reduce burden, and easy to use

Pending Publication Date: 2022-06-24
哈尔滨工业大学人工智能研究院有限公司
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] (1) The excellent performance of deep learning is rarely used in target tracking;
[0005] (2) Using the traditional method for block tracking requires a large number of parameter settings, which is relatively more complicated

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
  • Block target tracking method on multi-domain convolutional neural network
  • Block target tracking method on multi-domain convolutional neural network
  • Block target tracking method on multi-domain convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0037] Please see attached figure 1 , the present invention provides a kind of technical scheme:

[0038] A block tracking system based on a multi-domain convolutional neural network includes a video uploading module 1 , an online video acquisition module 2 , a video tracking module 3 , a tracking result saving module 4 and a data server 5 .

[0039] The video uploading module 1 loads the offline video from the local server, and the online video acquisition module 2 uses the high-definition camera equipment to collect the video online; the video tracking module 3 selects the tracking object after receiving the transmitted video sequence, and then calls the block based on deep learning. The tracking algorithm performs object tracking, and the image saving module 4 saves the tracking result of each frame to the data server 5 in the form of a picture.

Embodiment 2

[0041] Please see attached image 3 , this embodiment provides a technical solution on the basis of Embodiment 1:

[0042] A block target tracking method on a multi-domain convolutional neural network, comprising the following steps:

[0043] ① When video tracking starts, first choose to capture video online or upload it to the local server;

[0044] ②If the video is uploaded locally, the video uploading module 1 loads the video from the server;

[0045] ③ If the target object is tracked online, the online video capture module 2 turns on the camera for video capture;

[0046] ④The first frame of video tracking module 3 selects the tracking target and chooses whether to perform block tracking

[0047] 5. If the follow-up video adopts block tracking, then the video tracking module 3 uses the MDNET-based block tracking method to track the target, otherwise the module only uses the MDNET method to track the target;

[0048] The video tracking algorithm of segmented tracking ba...

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 block target tracking method on a multi-domain convolutional neural network in the technical field of computer vision, which comprises a video uploading module, an online video acquisition module, a video tracking module, a tracking result storage module and a data server, and is characterized in that the video uploading module uploads a video to be analyzed in a local data server; the on-line video acquisition module selects a target object through a camera and performs on-line target tracking, the video tracking module selects a tracking method to track the object to be tracked, and the tracking result storage module stores a tracking result in a server; according to the method, excellent characteristics of deep learning and block tracking are fully exerted, meanwhile, diversity selection of videos can be carried out, and a tracking result can be automatically stored; the algorithm for carrying out block tracking through the MDNET not only can exert the excellent tracking performance of the MDNET, but also can improve the tracking performance by combining the characteristics of blocking resistance and the like of blocks.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a block target tracking method on a multi-domain convolutional neural network. Background technique [0002] With the rapid development of information technology, people can process image and video information with the help of computers. Computer vision is more and more widely used in production and life, and it has become one of the most popular research fields for researchers. At present, relying on computer vision technology, tasks such as automatic acquisition of pictures and videos can be realized, and processing such as detection, tracking and analysis can be performed. At present, with the development of computer vision, target tracking has been widely used in various fields, such as: intelligent transportation, unmanned driving, national defense and security, intelligent monitoring, medical diagnosis, human-computer interaction and many other fields. Object trac...

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): G06T7/223G06N3/04G06N3/08
CPCG06T7/223G06N3/08G06T2207/10016G06T2207/20081G06T2207/20084G06N3/045
Inventor 孙鑫田威李栋王伟
Owner 哈尔滨工业大学人工智能研究院有限公司
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