Spektral Özellikler ve MFCC Tabanlı Özellikleri Kullanan Klasik Makine Öğrenmesi Metotlarıyla PCG Parça Sınıflandırması

Abstract

Cardiovascular diseases are some of the most common diseases today. Congenital abnormalities, diseases caused by impaired heart rhythm, vascular occlusion, post-operation arrhythmias, heart attacks and irregularities in heart valves are some of the various cardiovascular diseases. Early recognition of them is very important for obtaining positive results in treatment. For this purpose, it is tried to diagnose and detect cardiovascular diseases by listening to the sounds coming from the heart. During the rhythmic work of the heart, the contraction and relaxation of the heart chambers and the filling and discharge of blood from the heart into the veins create the sounds that are identified with the heart. Among the characteristic sounds of the heart, there can be some sounds similar to rustling which are indicators of pathological conditions. These unexpected sounds, similar to rustling, are called heart murmurs. Phonocardiograph device is used to record these mechanical sounds via microphone. Heart sounds recordings captured by a phonocardiograph device are called phonocardiograms (PCGs). Expert physicians try to detect the heart murmurs by listening to the heart sounds and examining PCGs. Ambient noise, the squeak of the microphone, and the patient's breathing sounds are the factors that make this task more difficult and challenging. Computer-aided systems supported with machine learning, signal processing and artificial intelligence algorithms offer solutions to help physicians in this regard. In this study, detection of heart murmur from PCG frames was examined. PCG frames of equal length, obtained by fragmenting the PCG recordings into 1-second-long frames, were classified by widely used machine learning methods namely C4.5 decision tree, Naive Bayes, Support Vector Machines and k-nearest neighbor. To train those classifiers we used spectral features of PCG signals, averages of MFCC values and some refined features obtained from a deep learning model which was inputted MFCC values. At the end of this manuscript the accuracies of those machine learning methods were compared.

Exploiting Data-Efficient Image Transformer-Based Transfer Learning for Valvular Heart Diseases Detection

“…Database Features Classification technique Domain Accuracy (%) [5] The [6] PhysioNet/CinC Statistical, Frequency XGBoost ML 92.90 [7] PhysioNet/CinC Time, MFCCs and Statistical Neural network (NN) ML 93.33 [11] PhysioNet/CinC MFCC Decision tree ML 86.40 [12] PhysioNet/CinC MFCCs NN ML 92.00 [13] PhysioNet/CinC LPC, Entropy, MFCCs, DWT and PSD NN ML 91.50 [14] Private MFCC DNN DL 91.12 [15] PhysioNet/CinC MFCCs DNN DL 93.00 [16] The HSM database TQWT, EMD and Shannon energy RBF neural networks ML 98.48 [17] PhysioNet/CinC MFSCs SVM ML 92.00 [18] PhysioNet/CinC Gram polynomial and Fourier transform NN ML 94.00 [19] PASCAL DWT Hidden Markov Models ML 92.74 [20] PhysioNet/CinC Wavelet CNN DL 81.20 [21] PhysioNet/CinC Modified frequency slice wavelet transform CNN DL 94.00 [22] PhysioNet/CinC Frequency spectrum, Energy and Entropy SVM ML 88.00 [23] Private EMD and MFCCs SVM ML 91.00 [24] MIT heart sounds Frequency SVM ML 98.00 [25] PhysioNet/CinC Time, MFCC, DWT and Wavelet SVM ML 82.40 [26] The HSM database FMFE + MFCC SVM ML 99.00 [27] PhysioNet/CinC Spectral SVM ML 98.00 [28] The HSM database Time-frequency MCC ML 98.33 [29] PhysioNet/CinC Cochleagram MLP ML 95.00 [30] The HSM database Time-frequency magnitude and phase RF ML 95.12 [31] Private MFCC KNN ML 98.00 [32] Private EMD KNN ML 94.00 [33] PASCAL MFCCs KNN ML 97.00 [34] Private MFCCs KNN ML 92.60 [37] PhysioNet/CinC MFCCs LSTM DL 98.61 [38] PhysioNet/CinC Time, Frequency and Time-frequency DNN DL 92.60 [39] PhysioNet/CinC MFCC+ MFSC 2D-CNN DL 81.50 [40] PhysioNet/CinC Mean, Standard deviation and Power spectrum CNN DL 86.02 [41] PhysioNet/CinC Spectrogram, Mel-spectrogram and MFCCs CNN DL 86.05 [42] The HSM database Normalized signals CNN...…”

“…These studies have explored the use of efficient hand-crafted feature extraction techniques in combination with effective classifiers. Several methods have been proposed for the detection of VHD, including Mel-Frequency Cepstral Coefficients (MFCCs) [11]- [15], Tunable Q-factor Wavelet Transform (TQWT) [16], Mel Frequency Spectral Coefficients (MFSCs) [17], Gram polynomial [18] and Wavelet Transform (WT) [19]- [21]. In ad-dition, the study examined various machine learning classifiers, including the support vector machine (SVM) [22]- [27], multiclass composite classifiers (MCC) [28], Multilayer Perceptron (MLP) [29], Random Forest (RF) [30], and k-Nearest Neighbor (k-NN) [31]- [34].…”