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

GPU-based Alternating Hough Forest Real-Time Object Tracking Method

A technology of alternating Hough forest and target tracking, applied in the field of image processing, can solve the problems of background interference, occlusion and low real-time performance

Inactive Publication Date: 2018-06-29
XIAMEN UNIV
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a GPU-based Alternating Hough Forest real-time target tracking method that can achieve stable and robust real-time target tracking for the problems of illumination changes, background interference, occlusion and low real-time performance in target tracking.

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
  • GPU-based Alternating Hough Forest Real-Time Object Tracking Method
  • GPU-based Alternating Hough Forest Real-Time Object Tracking Method
  • GPU-based Alternating Hough Forest Real-Time Object Tracking Method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0055] Concrete implementation steps of the present invention are as follows:

[0056] Step 1, obtain the training sample set.

[0057] (1a) Select a video sequence M(m) to be tested from the standard video library 1 ,...,m N ).

[0058] (1b) Based on the video sequence M, extract the training sample set Where N is the total number of samples, and each training sample contains 32 feature channels.

[0059] (1b1) Extract the region containing only the target from the given image containing the class label, i.e. the target region of the image;

[0060] (1b2) Feature extraction is performed on the image target area, including Lab features, HOG features and LBP features, including a total of 32 feature channels;

[0061] (1b3) Compute the integral map of the feature image extracted from each target region;

[0062] (1b4) Use random sampling to randomly extract a set of 16×16 image blocks {P i =(I i ,c i , d i )}, P i is an image block, I i is the feature of the integr...

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

A GPU-Based Alternating Hough Forest Method for Real-time Object Tracking. 1) Extract the training sample set according to the video sequence to be tested; 2) Set the number of random trees and the maximum depth of the forest in the alternating Hough forest; 3) Assign different weights to the training samples; 4) Initialize the root of each random tree node; 5) construct alternate Hough forest; 6) adopt splitting strategy when node splits; 7) copy alternate Hough forest into the texture memory of GPU; 8) manually determine the target area and center of the first frame image in the video , and set the search radius; 9) Use the constructed Alternating Hough Forest to detect subsequent frames in the GPU to obtain a confidence map about the center position of the target; 10) Copy the confidence map into the CPU memory; 11) Use the confidence map 12) Repeat steps 9 and 10 until completing the target tracking of all frames in the video sequence.

Description

technical field [0001] The invention relates to image processing, in particular to a GPU-based alternate Hough forest real-time target tracking method that can be used in the fields of intelligent monitoring, target tracking and human-computer interaction. Background technique [0002] Visual object tracking is an important and complex research content in the field of computer vision. After years of research, it has become one of the research hotspots in the field of computer vision. Visual object tracking refers to locating a specific object in a video sequence, obtaining its motion parameters, and then estimating its pose. Target tracking has a very wide range of applications in many fields, including automatic target detection, target monitoring, target activity analysis and human-computer interaction, and is widely used in factories, schools, transportation, hospitals, banks and other places. The intelligent video surveillance system is based on target tracking, and aft...

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): G06K9/00G06K9/62
CPCG06V20/40G06F18/24
Inventor 戴平阳游乔贝韩少华谢怡
Owner XIAMEN UNIV
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