Lung sound classification method, system and application based on convolutional neural network

Abstract

The present invention relates to a lung sound classification method, system and application based on a convolutional neural network. An image identification convolutional neural network is applied tolung sound classification, a general model is used to solve huge defects that a traditional algorithm has manual extraction of features and join experience limitation conditions and cannot be generally suitable for all the samples, and solve problems that a traditional algorithm cannot be enhanced with a data set and cannot be subjected to autonomic learning enhancement.

Description

technical field [0001] The present invention relates to lung sound classification technology, in particular to a convolutional neural network-based lung sound classification method, system and application. Background technique [0002] Over the past 40 years, work in this area has demonstrated limited success in identifying lung sounds: most published studies have used only a small number of patients (usually N<20), or focused on a single type of Lung sound; of course, very good classification results can be achieved, all algorithms (HMM, neural network, SVM vector machine, etc.) To achieve experimental results, a custom design and careful fit were performed to match the data. However, as the number of patients increases to tens to hundreds, features learned from small datasets may not be engineerable; based on our review of the lung literature over the past 40 years, the median number of these published studies was around 15 ; Despite advances in computing and algorith...

AI Technical Summary

Problems solved by technology

[0003] In view of the above problems and deficiencies, the present invention proposes a lung sound classification method, system and application based on a convolutional neural network. The image recognition convolutional neural network is applied to the classification of lung sounds, and a general model is used to solve the problem of manual extraction by traditional algorithms. Special, it is believed that the huge defects of adding restrictive conditions and not being universally applicable to all samples solve the problem that traditional algorithms cannot learn independently with the enhancement of data sets

Images