Fast Target Tracking Method Based on Recurrent Regression Network

A target tracking and basic network technology, applied in the field of fast target tracking based on cyclic regression network, can solve the problems of limiting the application scope and application effect of the target tracking method, affecting performance, etc., and achieve the effect of accurate positioning of the target and good robustness

Active Publication Date: 2021-01-29
HARBIN INST OF TECH
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, poor target tracking results will directly affect the performance of the above-mentioned related applications based on target tracking, thus limiting the application scope and application effect of the target tracking method to a certain extent.

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
  • Fast Target Tracking Method Based on Recurrent Regression Network
  • Fast Target Tracking Method Based on Recurrent Regression Network
  • Fast Target Tracking Method Based on Recurrent Regression Network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0016] The technical solution of the present invention will be further described below in conjunction with the accompanying drawings, but it is not limited thereto. Any modification or equivalent replacement of the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention should be covered by the present invention. within the scope of protection.

[0017] The present invention provides a fast target tracking method based on cyclic regression network, figure 1 Shown is the overall structure of the entire network, which can be roughly divided into three parts, as follows:

[0018] The first part is the basic network. At present, most tracking algorithms mostly use relatively lightweight networks such as VGG-M as the basic network. The reason why such a lightweight network is selected as the basic network is to be able to change drastically in the appearance of the target. Timely online fine-tuning,...

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 fast target tracking method based on a cyclic regression network. The method comprises the following steps: step 1, using the ResNet50 network as the basic network of the regression network; step 2, after the regression network is trained, on the basis of Introduce the LSTM network to form the final cyclic regression network to capture the various appearance changes of the target during the tracking process; step 3, use the Smooth-L1 loss function to train the cyclic regression network. The whole process of the present invention uses a neural network to track the target, uses deep supervision to return the coordinates of the target frame on features of different scales, and uses long and short-term memory networks to capture various appearance changes of the target during the tracking process. Compared with the existing target tracking method, it can locate the target more accurately without online update, and has good robustness.

Description

technical field [0001] The invention relates to a target tracking method, in particular to a fast target tracking method based on a cyclic regression network. Background technique [0002] The purpose of object tracking is to automatically mark the object in subsequent frames by giving the initial frame label box. With the advancement of object tracking technology, it is playing a role in more and more fields, such as video surveillance, human-computer interaction and motion recognition, etc. However, poor target tracking results will directly affect the performance of the above-mentioned related applications based on target tracking, thus limiting the application scope and application effect of the target tracking method to a certain extent. In recent years, object tracking has achieved great success due to the application of convolutional neural networks in the field of computer vision. Contents of the invention [0003] In order to perform better target tracking, the ...

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 Patents(China)
IPC IPC(8): G06T7/20
CPCG06T7/20G06T2207/10016G06T2207/20081G06T2207/20084
Inventor 邬向前卜巍马丁
Owner HARBIN INST OF TECH
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