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

Specific target candidate box generating method based on gauss model

A Gaussian model and specific target technology, applied in the field of target detection, can solve the problems of poor detection effect and low detection efficiency, and achieve the effect of improving target detection efficiency and detection effect, large auxiliary function, and improving detection efficiency

Inactive Publication Date: 2016-08-24
CHONGQING UNIV
View PDF4 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] This application solves the technical problems of low detection efficiency and poor detection effect in the prior art by providing a method for generating specific target candidate frames based on a Gaussian model

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
  • Specific target candidate box generating method based on gauss model
  • Specific target candidate box generating method based on gauss model
  • Specific target candidate box generating method based on gauss model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0040] A specific target candidate box generation method based on Gaussian model, such as figure 1 shown, including the following steps:

[0041] S1: Establish a Gaussian model at each pixel with the time-varying target scale as an independent variable, and initialize the Gaussian model mean μ and standard deviation σ at each pixel. In this embodiment, the Gaussian model mean μ=900, the Gaussian model Standard deviation σ=20;

[0042] S2: Train the Gaussian model, and dynamically filter out the location where the target frequently appears and the target scale information during the learning process to obtain the detection area;

[0043] S21: Set the standard deviation threshold σ according to the characteristic that the standard deviation of the Gaussian model reflects the peak value of the probability of the Gaussian curve t , the standard deviation threshold σ in this example t = 18, according to the mean value of the Gaussian model reflecting the characteristics of the p...

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 provides a specific target candidate box generating method based on gauss models. The method includes the steps of establishing gauss models of each pixel point with a target scale varying with time as an independent variable, training the gauss models, dynamically screening out the position where the target appears frequently and the target scale information in a learning process to obtain a detection area, and uniformly setting a candidate box in the detection area through a sliding window searching means according to the target scale information obtained during the gauss model learning. The gauss models are trained by using the inherent characteristics of special scenes inlcuding stationary monitoring videos and vehicle-mounted monitoring videos to obtain the area where the target appears frequently and the scale information of a detected point. On-line learning mechanism is employed, so that updating and detection of the model are conducted synchronously. The target detection efficiency and detection effect are substantially improved, and assistance for a subsequent specific target identifying and tracking process is provided.

Description

technical field [0001] The invention relates to the technical field of target detection, in particular to a method for generating a specific target candidate frame based on a Gaussian model. Background technique [0002] The generation of specific target candidate boxes is an important preliminary work for target recognition classification, and the generation of candidate boxes has a great impact on the accuracy and efficiency of the final specific target detection. If the candidate frames generated in the early stage have a large coverage with the real target, and the number of candidate frames is not enough to have a significant impact on the overall retrieval efficiency, it will have a positive effect on the entire target detection system, that is, to improve the accuracy of target detection. , and shorten the search time; in addition, the generation of target candidate boxes will also change the data distribution processed by the classifier, which may improve the detecti...

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): G06K9/62
CPCG06F18/214
Inventor 覃剑王美华肖婷张媛
Owner CHONGQING 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