Image recognizing method and system based on neural network

Abstract

The invention relates to an image recognizing method and system based on a neural network. The method includes the steps of firstly, collecting images of training samples and types of the training samples to set up a training sample set; secondly, setting original values of preset neural network parameters; thirdly, training the neural network according to the training sample set; fourthly, recognizing and classifying the images to be classified through the trained neural network, wherein the neural network comprises an input layer, an intermediate variable layer and an output layer, nodes of the intermediate variable layer include excitation type variable nodes of each output nerve cell node and suppression type variable nodes of each output nerve cell node, each node of the intermediate variable layer is connected with one of input nerve cell nodes of the input layer through a variable weight, and the variable weights include variable long-term weights and variable short-term weights. The required calculation amount from the input layer to the output layer is in direct proportion to the number of times of inputting the samples, namely, the calculation amount grows in a linear mode, and due to the method and system, the calculation amount is greatly reduced, and the recognition efficiency is improved.

Description

technical field [0001] The invention relates to the field of image recognition, in particular to an image recognition method and system based on a neural network. Background technique [0002] Unsupervised learning method is a kind of machine learning method, which is widely used in many research fields, such as data mining, computer vision, artificial intelligence, etc. For example, clustering methods such as K-mean can allow computers to automatically find regular data and remove irregular noise data. The K-mean clustering method can also be applied to the image field to classify and identify images. However, the calculation amount of the K-mean clustering method is proportional to the square of the number n of input samples, that is, the calculation amount is n 2 increased, resulting in low recognition efficiency. Contents of the invention [0003] Based on this, it is necessary to provide a neural network-based image recognition method that can improve the recognitio...

AI Technical Summary

Problems solved by technology

[0003] Based on this, it is necessary to provide a neural network-based image recognition method that can improve recognition efficiency for the problem of low image recognition efficiency in the prior art.

Images