Statistical feature embedding for heart sound classification

Abstract

Cardiovascular Disease (CVD) is considered as one of the principal causes of death in the world. Over recent years, this field of study has attracted researchers' attention to investigate heart sounds' patterns for disease diagnostics. In this study, an approach is proposed for normal/abnormal heart sound classification on the Physionet challenge 2016 dataset. For the first time, a fixed length feature vector; called i-vector; is extracted from each heart sound using Mel Frequency Cepstral Coefficient (MFCC) features. Afterwards, Principal Component Analysis (PCA) transform and Variational Autoencoder (VAE) are applied on the ivector to achieve dimension reduction. Eventually, the reduced size vector is fed to Gaussian Mixture Models (GMMs) and Support Vector Machine (SVM) for classification purpose. Experimental results demonstrate the proposed method could achieve a performance improvement of 16% based on Modified Accuracy (MAcc) compared with the baseline system on the Physionet2016 dataset.

A Comprehensive Survey on Heart Sound Analysis in the Deep Learning Era

“…Frequency which leads to the maximum spectrum [87], [88] Dominant frequency ratio Ratio of the maximun energy to the total energy [87], [88] Energy Spectral energy [38] Spectral roll-off Frequency below a specific percentage of the total spectral energy [32] Spectral centroid Average of magnitude spectrogram at each frame [32], [89] Specrtal flux Changing speed of the power spectrum [32] Power spectral density (PSD) Distribution of power in spectral components [46], [85], [89] Spectral entropy Shannon entropy of PSD [54], [86] Discrete cosine transform of Mel-scaled spectrogram [38], [75], [79], [86], [87], [93]- [95] Fractional Fourier transform-based Mel-frequency Mel-frequency from the fractional Fourier transform [43] Wavelet Wavelet transform Frequency analysis of a signal at various scales [59], [85], [89] Wavelet scattering transform "Wavelet convolution with nonlinear modulus and averaging scaling function" a (translation invariance and elastic deformation stability [96]) [96], [97] Wavelet synchrosqueezing transform Reassignment of wavelet coefficients [35] Tunable quality wavelet transform "Wavelet multiresolution analysis with a user-specified Q-factor, which is the ratio of the centre frequency to the bandwidth of the filters" b…”

“…The Naïve Bayes Classifier [41], [91], [102], [103] was widely used for heart sound classification due to its advantage of being easy-touse. Gaussian Mixture Models (GMMs) [53], [95] were used to estimate the data distribution by optimising the weights of Gaussian mixture components and mean and variance in each component. A Gaussian mixture-based HMM [38] was employed for heart sound classification considering the four sequential heart states, i. e., S1, systole, S2, and diastole.…”

“…4 presents a statistic of recent works from 2017 to 2022 that employ classic machine learning models for heart sound classification. SVMs have been very widely used for heart sound classification by learning a supporting hyperplane between classes [6], [33], [40], [41], [43], [46]- [48], [55], [60], [64], [75], [79], [85], [91]- [93], [95]- [97], [99], [100], [103]. Apart from linear projection between data samples and labels, SVMs can learn separating hyperplanes on non-linear data via non-linear kernels, such as radial basis function.…”

Unsupervised Heart Abnormality Detection Based on Phonocardiogram Analysis with Beta Variational Auto-Encoders

“…Traditional methods, such as support vector machine [5], i-vector [6] and hidden Markov model [7] are used to detect anomaly PCG signals in a supervised way. Deep learning methods based on Variational Auto-Encoder (VAE) [8], deep arXiv:2101.05443v1 [cs.SD] 14 Jan 2021 convolutional neural network [9,10] and recurrent neural network [11] are also used recently.…”

Unsupervised Phonocardiogram Analysis With Distribution Density Based Variational Auto-Encoders

“…With labels of normal and anomaly PCG signals, the PCG analysis can be considered as a classification problem. Classical machine learning techniques such as Support Vector Machine (SVM) ( 2 ), i-vector based dictionary learning method ( 3 ) and solutions based on Markov models ( 4 ) are used to solve the proposed problem besides deep learning algorithms ( 5 , 6 ). However, as a supervised problem, PCG data collected needs to cover all types of PCG abnormality, which is labor expensive.…”