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

Target tracking method with self-recovery force based on multistage feature extractor

A feature extraction and target tracking technology, applied in the field of computer vision, can solve the problems of affecting target tracking and low detection rate of algorithms, and achieve the effects of fast target tracking speed, strong robustness, and high target recognition accuracy

Inactive Publication Date: 2022-08-09
苏州海裕鸿智能科技有限公司
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In practical applications, since the tracked object and the tracking device are always in the process of relative motion, the scale of the tracked target in the collected image is always changing, and the drastic change usually leads to a change in the detection rate of the algorithm. Low, affects 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
  • Target tracking method with self-recovery force based on multistage feature extractor

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0010] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0011] figure 1 A self-resilient object tracking method based on multi-level feature extractor is shown, which utilizes feature extraction network (VGG) and multi-scale feature pyramid network to extract features from input images or videos, and then generates features based on the learned features. Dense bounding boxes and classification scores are subjected to non-maximum suppression operations to generate final results.

[0012] Specifically, the feature extraction network is composed of a feature fusion module, a U-shaped module and a multi-scale feature recombination network; the feature fusion module is composed of a feature fusion module 1 and a feature fusion module 2, and the feature fusion module 1 through the fusion backbone The feature map enriches semantic information as basic features (VGG); the U-shaped module has three groups, ...

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 discloses a multi-stage feature extractor-based target tracking method with self-restoring force, which is characterized by comprising the following steps of: extracting features from an input image or video by using a feature extraction network and a multi-scale feature pyramid network; and generating a dense bounding box and a classification score according to the learned features, and carrying out non-maximum suppression operation to generate a final result. According to the multi-level feature extractor-based target tracking method with the self-restoring force, the image feature pyramid is constructed by using the multi-scale features extracted by the multi-level feature extractor, so that the problem caused by the scale change of the tracking target is relieved, and the method has relatively high target recognition accuracy and relatively high target tracking speed; moreover, the method is high in robustness, has the self-recovery capability, and is higher in application value in the fields of target tracking, video monitoring and the like.

Description

technical field [0001] The invention relates to the field of computer vision, in particular to a target tracking method with self-resilience based on a multi-level feature extractor. Background technique [0002] Target tracking is a research hotspot in the field of computer vision, and it has high application value in the fields of cloud tracking, UAV, intelligent target tracking, and intelligent transportation. In real applications, since the tracked object and the tracking device are always in relative motion, the scale of the tracked target in the captured image is always changing, and the drastic change usually leads to a change in the detection rate of the algorithm. Low, affects target tracking. SUMMARY OF THE INVENTION [0003] The technical problem to be solved by the present invention is to provide a target tracking method with self-resilience based on a multi-level feature extractor, which can alleviate the problem caused by the change of the tracking target sc...

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): G06T7/246G06V10/25G06V10/80G06V10/82G06N3/04
CPCG06T7/246G06V10/25G06V10/806G06V10/82G06N3/04
Inventor 李璐胡浩锦邓维轩
Owner 苏州海裕鸿智能科技有限公司
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