Method for grey correlation analysis method to determine number of hidden layer characteristic graphs of convolutional neural network

A convolutional neural network and gray correlation technology, applied in the field of gray correlation analysis method to determine the number of hidden layer feature maps in the convolutional neural network network structure, can solve the problem of less design research on the number of hidden layer feature maps, CNN There are few discussions on the setting of network structure parameters, etc.

Active Publication Date: 2016-04-27
沈阳沙果科技有限公司
View PDF3 Cites 20 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] At present, researchers have used CNN for traffic sign recognition and achieved good results, but there are few discussions on the setting of CNN network structure parameters, especially the design research on the number of hidden layer feature maps in the network structure. few

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
  • Method for grey correlation analysis method to determine number of hidden layer characteristic graphs of convolutional neural network
  • Method for grey correlation analysis method to determine number of hidden layer characteristic graphs of convolutional neural network
  • Method for grey correlation analysis method to determine number of hidden layer characteristic graphs of convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0112] Embodiment: With reference to files 1 to 3, tables 1 to 5, a method for determining the number of hidden layer feature maps of a convolutional neural network using the gray relational analysis method, the steps are as follows:

[0113] (1) Download the standard traffic sign images collected by the German Institute of Neural Computing, and select 10 categories with more images, as shown in file 1.

[0114] (2) Convert the color image of file 1 into a grayscale image, perform histogram equalization processing on the traffic sign image, normalize it to a size of 28*28, and establish a grayscale image database of traffic signs, as shown in file 2.

[0115] (3) Then all images are binarized, and a binary image library of traffic signs is established, as shown in file 3.

[0116] (4) Set the number of layers of the network structure of the convolutional neural network and other related parameters: the size of the convolution kernel is 5*5; the number of layers of the convolut...

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 method for a grey correlation analysis method to determine the number of hidden layer characteristic graphs of a convolutional neural network, comprising steps of taking a binary image of a traffic sign as an image database, using the convolutional neural network method to perform identification, introducing a gray correlation analysis method in a network training process in order to automatically choose a hidden layer characteristic image which has a greatest impact on the identification result so as to optimize a network structure, taking the traffic sign as an object to perform identification and performing optimization on the CNN network structure. The experiment shows that the number of the characteristic image can be adaptively determined by using the method and the optimization of the CNN network structure can be finished. Compared with the experiment method, the invention can improve and determine the efficiency of the network.

Description

Technical field: [0001] The invention relates to a method for determining the number of hidden layer feature maps in a convolutional neural network network structure, in particular to a gray correlation analysis method for determining the number of hidden layer feature maps in a convolutional neural network network structure. Background technique: [0002] Convolutional neural network is a kind of artificial neural network. It is a new type of neural network based on multi-layer supervised learning network. It has become a hot spot in the fields of speech analysis and image recognition. Due to its weight sharing feature, it reduces the complexity of the network model and reduces the number of weights. And the convolutional neural network integrates the feature extraction function into the classifier, omitting the complex feature extraction process before recognition, so it is widely used in image recognition, object detection and recognition, and target tracking. [0003] A...

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): G06N3/08G06K9/00
CPCG06N3/08G06N3/082G06N3/084G06V20/582
Inventor 张志佳李媛媛唐岩钟玲于雅洁
Owner 沈阳沙果科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap