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

Level set target tracking method based on convolutional neural network

A convolutional neural network and target tracking technology, which is applied in image data processing, instrumentation, computing, etc., can solve the problems of low efficiency of tracking module algorithm, inability to obtain target outline, target detection, insufficient efficiency and accuracy of outline extraction, etc.

Inactive Publication Date: 2018-02-16
BEIJING INSTITUTE OF TECHNOLOGYGY
View PDF7 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The system can obtain the tracking frame of the tracking target, but only the position and size of the target, not the specific outline of the target
[0010] The above-mentioned existing tracking algorithms have completed the task of target tracking to a certain extent, however, they have shortcomings in the efficiency and accuracy of target detection and contour extraction.
The present invention aims to solve the problem that the existing target tracking algorithm is not sensitive to the deep learning features of the video frame and the tracking module algorithm has low efficiency, and proposes a more efficient method based on convolutional neural network and level set model that can be used for long-term tracking Object Tracking Method

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
  • Level set target tracking method based on convolutional neural network
  • Level set target tracking method based on convolutional neural network
  • Level set target tracking method based on convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0082] This embodiment illustrates the application of the present invention "a level set target tracking method based on convolutional neural network" to the video "cheetah" (available from the database: http: / / cpl.cc.gatech.edu / projects / SegTrack / The process of a certain frame in download):

[0083] figure 1 For the algorithm flow of this method and this embodiment, from figure 1 It can be seen that the method includes the following steps:

[0084] Step A: Initialize the tracking program;

[0085] Specifically in this embodiment, the input video frame I 1 For the first frame of "cheetah", the input contour C 1For the first frame standard segmentation result provided by the database, extract features and obtain Adaboost strong classifier. The features are divided into three parts:

[0086] 6 different channel features f of RGB (Red, Green, Blue) and HSV (Hue, Saturation, Value) color spaces 1 -f 6 , a total of 6 dimensions (n color = 6);

[0087] HOG feature f for 5×...

Embodiment 2

[0099] This embodiment specifically illustrates the tracking results obtained by executing steps 1 to 4 of the present invention on two video frames "cheetah" and "monkeydog" (the tracking outline is represented by a white line).

[0100] figure 2 Divided into 2 rows and 3 columns, each row represents a different frame of a video, and 3 columns represent the tracking results of the middle three frames of the video;

[0101] From figure 2 It can be seen from the figure that this method can accurately capture the outline of objects on a series of images and maintain high-quality tracking. The validity of the tracking method is illustrated.

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 relates to a level set target tracking method based on a convolutional neural network, and belongs to the technical field of target tracking and video processing. The level set target tracking method based on a convolutional neural network includes the following steps: 1) utilizing the first frame and the first frame standard target contour of video input to initialize an Adaboost detection module; 2) calling a detection module to obtain the initial position and shape of an object in the subsequent video frame; 3) operating a level set method tracking module on the above basis, and accurately segmenting the object contour; and 4) by means of the result of the step 3, distinguishing the foreground / background, and updating a weak classifier of the detection module. For the level set target tracking method based on a convolutional neural network, the utilization method of convolutional neural network characteristics can make a preferable distinction on the foreground / background so as to improve the detection accuracy and improve the overall target tracking effect; and the level set target tracking method based on a convolutional neural network can balance performance andefficiency, and can obtain relatively excellent tracking effect at the same operating speed.

Description

technical field [0001] The invention relates to a level set target tracking method based on a convolutional neural network, and belongs to the technical fields of target tracking and video processing. Background technique [0002] Object tracking is a widely used technique. In the field of computer video processing, the target tracking problem generally refers to calculating the position or contour of the object in each subsequent frame given the position or contour information of the first frame in the video sequence. Object tracking remains a challenging problem due to problems such as deformation, occlusion, and scale transformation of objects in video sequences. In order to solve this problem, the TLD (Tracking-learning-detection) algorithm combines tracking and detection (Detection), laying the foundation for the tracking by detection (Tracking by Detection) algorithm. At present, more efficient algorithms are generally developed under this framework. They are mainly ...

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
IPC IPC(8): G06T7/246
CPCG06T7/246G06T2207/10016G06T2207/20081G06T2207/20084
Inventor 刘利雄宁小东
Owner BEIJING INSTITUTE OF TECHNOLOGYGY
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