An image classification method based on multi-layer spiking convolutional neural network

A technology of convolutional neural network and classification method, applied in the field of image classification based on multi-layer spiking convolutional neural network, can solve the problems of non-convergence of learning and voltage redundancy, so as to reduce the number of total pulses, ensure convergence, and reduce calculation The effect of complexity

Active Publication Date: 2020-12-01
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The object of the present invention is: the present invention provides a kind of image classification method based on multi-layer spiking convolutional neural network, solves the technical problem of the redundant calculation of spiking neuron membrane voltage due to the increase of neuron scale, solves simultaneously The SNN activation function is not derivable, and it is impossible to use backpropagation to calculate the residual, which leads to technical problems that the learning does not converge

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
  • An image classification method based on multi-layer spiking convolutional neural network
  • An image classification method based on multi-layer spiking convolutional neural network
  • An image classification method based on multi-layer spiking convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0047] A kind of image classification method based on multilayer spiking convolutional neural network, comprises the following steps

[0048] Step 1: Convert the image in the training set to a pulse sequence, that is, image preprocessing, use the contrast coding method to strengthen the edge information of the image and convert it into a pulse sequence;

[0049] The step 1 is specifically:

[0050] Step 11: The input is the MNIST digital handwriting image data set, the size of the image is 28*28, and the upper bound of the pixel distance of the image is set to d=1 and the maximum time T of neuron pulse emission max =100ms, define the image matrix as A, the pixel value matrix as pixel, both A and pixel are initialized to a 28*28 matrix, and each pixel in the image is p;

[0051] Step 12: Calculate the Euclidean distance of the pixel point p in space, define the pixel point whose Euclidean distance is less than the upper bound d of the distance as q, and add it to the set Γ q ...

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 an image classification method based on a multi-layer spiking convolutional neural network, relates to the field of image processing, converts images in a training set into pulse sequences; sets the parameters of spiking neurons, and uses the spiking neurons to construct a convolutional neural network ; Using the pulse sequence as input, train the convolutional neural network layer by layer, obtain the classification result after obtaining the visual features of the pulse sequence, the training method is an unsupervised learning algorithm based on hebbian rules; the image to be identified Convert it into a pulse sequence, input the trained convolutional neural network, and get the classification result of the image to be recognized; this method solves the technical problem of redundant calculation of spiking neuron membrane voltage due to the increase of neuron scale, and solves the The activation function is not derivable, and it is impossible to use backpropagation to calculate the residual, which leads to technical problems that the learning does not converge.

Description

technical field [0001] The invention relates to the field of image processing, in particular to an image classification method based on a multi-layer spiking convolutional neural network. Background technique [0002] Image classification is an image processing method that distinguishes different types of objects according to the different characteristics reflected in the image information. It uses computer to carry out quantitative analysis on images, and classifies each pixel or area in the image or image into one of several categories to replace human visual interpretation. [0003] Spiking neural network (SNN) is a mathematical model based on pulse timing that highly fits biological neural networks, and simulates the biological electrical response of biological neurons through the triggering of pulses. Each neuron is an independent calculation and response unit, which is connected through a large-scale parallel network to form a powerful mathematical model and is used t...

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 Patents(China)
IPC IPC(8): G06K9/62G06N3/04
CPCG06N3/045G06F18/24
Inventor 屈鸿秦展展柏晓祉张亚肖艳清
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA
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