A Method of Video Target Tracking

A target tracking and video technology, which is applied in the field of image and video target tracking, can solve the problems of slow speed of particle filter algorithm, and achieve the effects of fast tracking, speed improvement and real-time guarantee

Active Publication Date: 2022-03-29
DALIAN UNIV OF TECH
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Secondly, the traditional particle filter needs to judge each image block, and in order to maintain the accuracy, there must be enough particles, resulting in a very slow particle filter algorithm.

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 Method of Video Target Tracking
  • A Method of Video Target Tracking
  • A Method of Video Target Tracking

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0023] The specific implementation manners of the present invention will be further described below in conjunction with the technical solutions.

[0024] A method for video target tracking, the steps are as follows:

[0025] 1. Offline training phase:

[0026] Step 1: Use the public database to train a classifier network. The input of the classifier network is an image block, and the output is the score of the image block. The output is 1 if the image block is the target foreground, and the output is 0 if the image block is the background; classification The detector network is used to distinguish whether the input image block is the target foreground or the background, and to score the image block; such as formula (1), where, Indicates the i-th image block obtained in the t-th frame of the video sequence, representing image blocks In the feature of the mth class in the classifier network 1, m takes the value of 1, 2; Represents the output of the mth class in the softm...

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 belongs to the technical field of image and video target tracking, and provides a video target tracking method, which can continuously track a specific single target in a video, and involves related knowledge of image processing. First, we train a fast object tracker using deep mutual learning and knowledge distillation. Secondly, when each frame comes, a lot of particles are scattered around the previous frame, and the distribution of particles is random. Then we select a large image region that can contain all particles. The relative position of the image area and particles is sent to the target tracker to get the final score, and the result with the highest score is selected. The final result is taken as the final result after going through the bounding box regression. Finally, the tracker is updated online every time tracking fails or after a certain amount of time. The benefit of the present invention is that the traditional sampling method is changed, and the sampling at the image layer is changed to the sampling at the feature layer, which greatly improves the speed and improves the speed under the condition of ensuring the accuracy.

Description

technical field [0001] The invention belongs to the technical field of image and video target tracking, can continuously track a specific single target in a video, and involves related knowledge of image processing. Background technique [0002] With the continuous development of image processing technology, video object tracking, because of its practicality, plays an important role in daily life. [0003] Video object tracking is mainly divided into two categories: particle filter methods and correlation filter methods. The correlation filter method is to use the target features to perform correlation matching around the target in the previous frame, and the final corresponding highest place obtained is the target position in the current frame. This method can convert the calculation to the Fourier domain because of the circular matrix method proposed in the 'High-speed tracking with kernelized correlation filters' published in the PAMI journal by J.F.Henriques et al. in 2...

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): G06T7/246G06V10/764G06K9/62
CPCG06T7/246G06T2207/10016G06T2207/20081G06F18/24
Inventor 卢湖川高凯
Owner DALIAN UNIV 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