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

Single target tracking system and method based on attention-key point prediction model

A prediction model and key point technology, applied in image processing, can solve the problems that the tracker cannot continue to track the target, cannot adapt to large-scale changes of the target, and the tracker is affected

Pending Publication Date: 2021-09-10
CHONGQING UNIV OF POSTS & TELECOMM
View PDF13 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The current tracking algorithm has been able to achieve precise positioning without interference, but if there are many similar targets around, the tracker will be greatly affected, causing the key point to be located on a similar target
Secondly, the current tracking algorithm cannot adapt to the large-scale changes of the target. When the target undergoes large-scale deformation or irreversible deformation, the tracker cannot continue to track the target.

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
  • Single target tracking system and method based on attention-key point prediction model
  • Single target tracking system and method based on attention-key point prediction model
  • Single target tracking system and method based on attention-key point prediction model

Examples

Experimental program
Comparison scheme
Effect test

no. 2 example

[0047] 1. Target positioning module:

[0048] 1. Label processing: For the first frame of image, Gaussian processing is performed on the two key points of the upper left corner and the lower right corner of the bounding_box, with the key point as the center and a circle with a radius of r as the key point candidate area, except for these two The key point area, other areas on the picture are uniformly marked as 0.

[0049] 2. Data augmentation processing: Do data augmentation processing on the first frame image, input the generated image into the double-layer convolutional neural network for training, use the conjugate gradient descent algorithm, and perform m (m<=5) times update operation,

[0050] 3. Predict the next frame: input the next frame picture into the model, output a score_map, then use the two largest points on the score_map as key points, after p (p<=10) frames, again Updates are performed using the conjugate gradient algorithm until the end of the last frame. ...

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 requests to protect a single target tracking system and method based on an attention-key point prediction model, and the system comprises a target state estimation module and a target positioning module, in the target positioning module, the position of a target is determined through the prediction of two key points, and the key points refer to the two corner points of the upper left corner and the lower right corner of bounding_box. In the target state estimation module, an attention mechanism is added to improve the discrimination capability and robustness of the model, and the complete state of the target is clearly marked by using bounding_box.

Description

technical field [0001] The invention belongs to the technical fields of deep learning, image processing, and target tracking, and especially adds the detection and attention mechanism of key corners to the target positioning and target state estimation modules, which enhances the discriminative ability and robustness of the model. Background technique [0002] In the context of single object tracking, it is often necessary to separate object localization and object state estimation into two independent but related subtasks. Object localization is basically to determine that the object to be tracked is located at a certain position in the image. However, the target localization module can only obtain position information, that is, the coordinate state of the target in the image, and the purpose of target state estimation is to find the complete state. In recent years, many famous foreign researchers have successfully solved this problem by online training powerful classifier...

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): G06T7/246G06T7/11G06N3/04G06N3/08
CPCG06T7/246G06T7/11G06N3/08G06T2207/20081G06T2207/20084G06T2207/10016G06N3/045Y02T10/40
Inventor 孙开伟王支浩郭豪邓名新刘期烈
Owner CHONGQING UNIV OF POSTS & TELECOMM
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