Image recognition and classification method based on prediction increment width learning

Abstract

The invention discloses an image recognition and classification method based on prediction increment width learning, and the method comprises the following steps: 1, building a width learning model, inputting the initial number of various types of nodes, and setting the recognition accuracy; 2, calculating the number of required enhancement nodes through a fitting function; 3, increasing the number of enhancement nodes, and carrying out incremental width learning; 4, judging whether the test identification accuracy is greater than the set identification accuracy or not, and if not, returning to the step 3, and if yes, outputting a training result and recognition accuracy. Partial experimental data of width learning is fitted through a least square method, the relationship between the number of enhancement nodes and the recognition precision is obtained through partial experimental data fitting, and the proper number of enhancement nodes meeting the precision can be obtained by settingthe precision, so the tedious steps of manually setting model parameters and the increase of training time caused by setting improper model parameters are avoided, and the method has the advantages ofhigh recognition precision and simple algorithm.

Description

Technical field[0001]The present invention relates to the field of image recognition, in particular to an image recognition classification method based on prediction increment width learning.Background technique[0002]With the continuous development of the Internet, pictures have gradually become the main carrier of dissemination of information, so the processing of images is very important, and image recognition technology as an important part of image processing has continued to develop in recent years. Machine learning is now the main method of image recognition technology, which learns and updates parameters through a large amount of external data to make recognition judgments. As the current mainstream of machine learning, deep learning has achieved good results in image processing. However, deep learning generally has a common problem, that is, many parameters need to be calculated and training time is long.[0003]The document "Broad Learning System: An Effective and Efficient I...

AI Technical Summary

Problems solved by technology

In order to improve the recognition accuracy of the width learning system, it is necessary to modify the parameters in the model. The change of the model is to manually set the number of various nodes, and increasing the fitting ability of the model by adding a large number of enhanced nodes will increase the entire training time.

Images