基于改进的MFCC与CNN心音信号识别方法的研究  

Research on Recognition Method of Heart Sound Signals Based on Improved MFCC and CNN

作　　者：王佳佳 熊飞龙 WANG Jiajia;XIONG Feilong(School of Energy and Mechanical Engineering,Jiangxi University of Science and Technology,Nanchang 330000,China)

机构地区：[1]江西理工大学能源与机械工程学院,南昌330000

出　　处：《计算机测量与控制》2024年第12期201-207,215,共8页Computer Measurement &Control

基　　金：江西省教育厅项目(GJJ21084)。

摘　　要：心音分类在心血管疾病的早期检测中起着至关重要的作用,特别是对小型初级卫生保健诊所、缺少专业人员陪护的家庭等检测;为提高心音信号数据类别间的可辨别性,提出了一种改进MFCC方法提取数据特征,并与PCA算法组合,作为样本输入CNN模型进行分类;对心音信号数据集进行降噪与下采样,减少数据量及噪声影响,利用改进的MFCC对其进行特征提取,并利用PCA算法进而抽取相关特征;为验证不同特征数据集以及不同滤波算法在提取心音数据特征数据集方面对分类过程及分类结果所产生的影响,将其分别输入CNN模型进行训练;经实验验证,改进的MFCC特征+PCA算法与传统的MFCC相比较,可提高训练模型的训练速度,同时也可提高识别率。Heart sound classification plays a crucial role in the early detection of cardiovascular disease,especially in small primary health care clinics and families lacking professional care.In order to improve the distinguishability between the data categories of heart sound signals,an improved Mel-Frequency Cipstal Coefficients(MFCC)method is proposed to extract data features and combine with Principal Component Analysis(PCA)algorithm,with a sample input Convolution Neural Network(CNN)model for classification.The heart sound signal data set is denoised and downsampled to reduce the amount of data and the influence of noise,their features are extracted by using the improved MFCC,and then,the relevant features by using the PCA algorithm.,In order to verify the effects of different feature data sets and filtering algorithms on the classification process and results in the heart sound data feature dataset extraction,they are input into the CNN model for training respectively.Experimental results show that compared with the traditional MFCC,the improved MFCC feature+PCA algorithm can improve the training speed and recognition rate of the training model.

关 键 词：心音分类 MFCC 滤波 PCA CNN 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]