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

A real-time tracking method for multi-channel kernel correlation filtering

A technology of kernel correlation filtering and real-time tracking, applied in the fields of image processing and computer vision, to achieve the effect of improving performance and speed

Active Publication Date: 2019-04-30
NANJING UNIV OF INFORMATION SCI & TECH
View PDF1 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The technical problem to be solved in the present invention is to fuse multi-channel features through kernel functions, extend linear correlation filtering to nonlinear correlation filtering, and propose a real-time tracking method for multi-channel kernel correlation filtering to ensure that the tracker can track in complex real-world scenarios. , such as occlusion, illumination changes, appearance changes, etc., it can still track the target accurately and quickly

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
  • A real-time tracking method for multi-channel kernel correlation filtering
  • A real-time tracking method for multi-channel kernel correlation filtering
  • A real-time tracking method for multi-channel kernel correlation filtering

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0042] In order to make the purpose, technical route and beneficial effects of the present invention clearer, the present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0043] Since the appearance of the target between two consecutive frames changes very little (high similarity), a filter template is obtained by learning the target information of the previous frame through ridge regression, and the image of the current frame is detected with the obtained filter template, and the corresponding The peak position in the filtered response is the target position for the current frame. The method of the present invention is mainly divided into three parts: the training stage, the detection stage and the update stage. The training stage: process the target information of the last frame through the ridge regression method to obtain the filter template; the detection stage: use the obtained filter template to filter th...

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 real-time tracking method of multi-channel kernel correlation filtering. The method includes: training stage: processing the target information of the last frame through the ridge regression method to obtain a filter template; detection stage: using the obtained filter template to process the current The image of the frame is detected, and the filter response is output; the update stage: the filter template and the target appearance are updated in real time. The method of the present invention utilizes the kernel function to fuse multi-channel features, and overcomes the selection limitation of multi-channel features. And through the kernel function, the linear optimization problem of ridge regression is converted into a nonlinear optimization problem in high-dimensional space, so as to construct a more robust filtering template to adapt to various scene changes of the target during the tracking process, and improve the tracking performance of the tracker. At the same time, bypassing the process of extracting a large number of samples and building complex appearance models, the speed of the tracker is greatly improved, which can meet the tracking needs of the real world.

Description

technical field [0001] The invention belongs to the field of image processing and computer vision, and in particular relates to a real-time tracking method of multi-channel kernel correlation filtering. Background technique [0002] In computer vision, object tracking is a fairly extensive research field, and it has a very wide range of applications in the fields of automatic monitoring, video indexing, traffic monitoring, and human-computer interaction. Although researchers have proposed many algorithms in the past decade, how to build a stable and efficient tracking system to deal with the appearance changes, fast motion, scale changes and occlusions of objects is still a challenging task. [0003] Most of the existing high-precision trackers build complex appearance models and extract a large number of candidate particles, and calculate the similarity or confidence value between each candidate particle and the tracking result of the previous frame through traversal. Ther...

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): G06K9/62
CPCG06F18/214
Inventor 胡昭华邢卫国王珏郭业才
Owner NANJING UNIV OF INFORMATION SCI & TECH
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