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

Target forecasting and tracking method based on recurrent neural network

A cyclic neural network and target prediction technology, applied in the field of visual tracking, can solve problems such as inability to fully respond, and achieve a good tracking effect.

Active Publication Date: 2018-07-06
FUZHOU UNIV
View PDF4 Cites 26 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, when the tracking tasks faced include various complex situations and interactions, such as internal changes in targets such as deformation, attitude changes, and complex motions, and external changes such as illumination changes, occlusions, and background clutter, existing methods cannot fully cope

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
  • Target forecasting and tracking method based on recurrent neural network
  • Target forecasting and tracking method based on recurrent neural network
  • Target forecasting and tracking method based on recurrent neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0047] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0048] like figure 1 , figure 2 as well as image 3 As shown, this embodiment provides a target prediction and tracking method based on a recurrent neural network, which specifically includes the following steps:

[0049] Step S1: Input the current frame into the single network of the feature extraction model SFE-Net+ to obtain deep features;

[0050] Step S2: Input the depth feature obtained in step S1 and the center position of the target in the previous frame into the motion model SM-Net, predict the center position of the target in the current frame, and randomly sample candidate frames around the center position;

[0051] Step S3: Input the candidate frame obtained in step S2 and the target frame of the first frame into the matching model for similarity determination one by one, select the candidate frame with the highest similarity, and obtain t...

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 target forecasting and tracking method based on a recurrent neural network. The method includes the steps: firstly, extracting characteristic information according to a current video frame, acquiring 1000-dimensional characteristics by a single network of an SFE-Net+ model and inputting the characteristics to a motion model SM-Net; secondly, forecasting the center position of a target in a current frame according to the characteristic information and the center position of a previous frame target by the SM-Net of a time structure; thirdly, selecting a certain numberof candidate boxes from the periphery of the center position in a random sampling mode; finally, inputting the candidate boxes to a matching model SMT-Net, judging similarities of the candidate boxesand target boxes of a first frame one by one, and selecting the candidate boxes with the highest similarity. An area occupied by the candidate boxes with the highest similarity in an original image isa final area of the target in the current frame. Objects can be tracked in complicated situations.

Description

technical field [0001] The invention relates to the field of visual tracking, in particular to a target prediction and tracking method based on a cyclic neural network. Background technique [0002] The core task of visual target tracking is to predict the trajectory of the target, extract the effective features of the target area, establish a suitable matching function, and complete the capture of target state information in the video sequence. In terms of medical diagnosis, the use of target tracking can effectively analyze the movement of organs and cells, and provide auxiliary diagnostic information for doctors; In terms of human-computer interaction, tracking technology provides a new non-contact and flexible way for humans to interact with computers, for example, by capturing the trajectory of gestures and identifying them to generate input information; in automatic In terms of driving, visual tracking is an important part of active and safe driving. It detects and tr...

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): G06T7/246G06N3/04
CPCG06T7/246G06N3/045
Inventor 余春艳林晖翔陈吕财郭文忠
Owner FUZHOU 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