Research on Recognition Method of Heart Sound Signals Based on Improved MFCC and CNN
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 ex-traction,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.
万方数据
维普
