Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Adaptive fusion complementary learning real-time tracking method based on target probability model

A probabilistic model, real-time tracking technology, applied in image analysis, instrumentation, calculation, etc., can solve problems such as target loss and damage tracker performance, and achieve the effect of excellent performance and versatility

Active Publication Date: 2022-07-15
FUJIAN INST OF RES ON THE STRUCTURE OF MATTER CHINESE ACAD OF SCI
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] In terms of multi-feature response fusion, most current tracking methods use fixed fusion coefficients, which will have different effects on feature fusion in different situations, and even seriously damage the performance of the tracker, resulting in target loss.

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
  • Adaptive fusion complementary learning real-time tracking method based on target probability model
  • Adaptive fusion complementary learning real-time tracking method based on target probability model
  • Adaptive fusion complementary learning real-time tracking method based on target probability model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0058] The present application will be described in detail below with reference to the examples, but the present application is not limited to these examples.

[0059] see figure 1 , the real-time tracking method for adaptive fusion and complementary learning based on the target probability model provided by this application includes the following steps:

[0060] Step S100: take the target position P in the t-1 frame image t-1 Be the center, take the t-1 frame image size as the matching area size, and generate the search area o in the multiple of the target size in the t-1 frame image;

[0061] Step S200: obtain the directional gradient histogram feature of the search area o and compare it with the t-1 frame image I t-1 The generated directional gradient histogram features are matched to obtain a second matching value matrix, and the second matching value matrix is ​​used as the directional gradient histogram matching value matrix r cf ;

[0062] Step S300: obtain the colo...

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 present application discloses an adaptive fusion complementary learning real-time tracking method based on a target probability model. Based on the current mainstream tracking method complementary real-time tracking (Staple), the method innovatively uses a piecewise function. When the foreground ratio of the tracking target obtained by the histogram feature is less than the segment function threshold, the average adaptive fusion coefficient is used. When the tracking target foreground ratio obtained by using the color histogram feature is greater than or equal to the segment function threshold, the exponential adaptive fusion coefficient is used. . Thereby, the appropriate segmentation function threshold can be selected according to different video attributes. This method has better performance and generality, and can be used to solve similar multi-feature fusion problems.

Description

technical field [0001] The present application relates to an adaptive fusion complementary learning real-time tracking method based on a target probability model, which belongs to the field of machine vision target tracking. Background technique [0002] Visual tracking technology is a hot and difficult point in the field of computer vision research. At the same time, it has broad prospects for commercial application and is widely used in human-computer interaction, public safety, medical imaging and other fields. [0003] At present, the commonly used target tracking methods use the directional gradient histogram of the target image and the color feature to achieve complementary advantages. Using the color feature to make up for the directional gradient histogram can only extract the target spatial information and must ignore the color feature. Although the directional gradient histogram makes up for the color feature feature, but only the target color information can be ex...

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): G06T7/246G06T7/90
Inventor 董秋杰周盛宗何雪东葛海燕
Owner FUJIAN INST OF RES ON THE STRUCTURE OF MATTER CHINESE ACAD OF SCI
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com