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

Correlation filtering tracking method based on background perception

A correlation filter and correlation filter technology, applied in image data processing, instruments, calculations, etc., can solve the problem of ignoring the background information of the tracking area, and achieve the effect of accurately predicting the target position

Active Publication Date: 2020-05-15
北京理工大学重庆创新中心 +1
View PDF9 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] However, most current methods only focus on foreground information, relying on handcrafted features such as HOG and CN, or deep features, ignoring background information in the tracked region

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
  • Correlation filtering tracking method based on background perception
  • Correlation filtering tracking method based on background perception
  • Correlation filtering tracking method based on background perception

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0025] The present invention will be further described below in conjunction with the accompanying drawings.

[0026] In the actual target tracking process, the background information in a given image sequence is often similar in consecutive frames, which helps to recognize a given target effectively. Therefore, we introduce a background-aware adjustment term, which can help us further mine the background information and learn a more robust correlation filter, which can be expressed as:

[0027]

[0028] in, is an area diagonal mask matrix, and its specific form is as figure 2 as shown, is the input sample feature, f∈R DK×1 is the final desired output correlation filter. The background-aware regularization term is added in the training process to achieve the purpose of giving similar importance to different regions of the image by balancing the output responses of each region of the image.

[0029] Based on this, if figure 1 As shown, a kind of correlation filtering...

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 provides a correlation filtering tracking method based on background perception, and the method comprises the following steps: step 1, reading the position information of a first frame image of a video and a to-be-tracked target, and adding the position information into a training set; step 2, extracting target area features in the training set to train a correlation filter with background perception; step 3, reading a next frame of video image and extracting image features of a search area; step 4, performing convolution operation on the extracted search area image features anda related filter to obtain current frame to-be-tracked target position information, and adding the current frame video image and the to-be-tracked target position information into a training set; andstep 5, repeating the steps 2-4 until all frames of the video are read to obtain the position information of the to-be-tracked target in each frame of video image. The background information in the tracking process can be effectively utilized, and the target position can be more accurately predicted.

Description

technical field [0001] The invention relates to the field of single target tracking, in particular to a correlation filter tracking method based on background perception, which is suitable for background perception problems in target tracking aspects such as intelligent monitoring and human-computer interaction. Background technique [0002] Video object tracking is one of the most important research directions in the field of computer vision, and is widely used in security, transportation, military and other research fields. The main task of single target tracking is to determine the position information of our tracking target in each frame of the video given the first frame of the video and the position frame of the tracking target. [0003] In recent years, visual object tracking methods based on correlation filters have developed rapidly and achieved remarkable results. The first tracker based on correlation filtering is MOSSE (Minimum Output Sum of Squared Error), whic...

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 Applications(China)
IPC IPC(8): G06T7/246G06T7/73
CPCG06T7/246G06T7/73G06T2207/10016G06T2207/20081G06T2207/20024
Inventor 许廷发张语珊吴零越吴凡
Owner 北京理工大学重庆创新中心