Heart Sound Segmentation-An Event Detection Approach Using Deep Recurrent Neural Networks

HBNET: A blended ensemble model for the detection of cardiovascular anomalies using phonocardiogram

BACKGROUND

Cardiac diseases are highly detrimental illnesses, responsible for approximately 32% of global mortality [1]. Early diagnosis and prompt treatment can reduce deaths caused by cardiac diseases. In paediatric patients, it is challenging for paediatricians to identify functional murmurs and pathological murmurs from heart sounds.

OBJECTIVE

The study intends to develop a novel blended ensemble model using hybrid deep learning models and softmax regression to classify adult, and paediatric heart sounds into five distinct classes, distinguishing itself as a groundbreaking work in this domain. Furthermore, the research aims to create a comprehensive 5-class paediatric phonocardiogram (PCG) dataset. The dataset includes two critical pathological classes, namely atrial septal defects and ventricular septal defects, along with functional murmurs, pathological and normal heart sounds.

METHODS

The work proposes a blended ensemble model (HbNet-Heartbeat Network) comprising two hybrid models, CNN-BiLSTM and CNN-LSTM, as base models and Softmax regression as meta-learner. HbNet leverages the strengths of base models and improves the overall PCG classification accuracy. Mel Frequency Cepstral Coefficients (MFCC) capture the crucial audio signal characteristics relevant to the classification. The amalgamation of these two deep learning structures enhances the precision and reliability of PCG classification, leading to improved diagnostic results.

RESULTS

The HbNet model exhibited excellent results with an average accuracy of 99.72% and sensitivity of 99.3% on an adult dataset, surpassing all the existing state-of-the-art works. The researchers have validated the reliability of the HbNet model by testing it on a real-time paediatric dataset. The paediatric model's accuracy is 86.5%. HbNet detected functional murmur with 100% precision.

CONCLUSION

The results indicate that the HbNet model exhibits a high level of efficacy in the early detection of cardiac disorders. Results also imply that HbNet has the potential to serve as a valuable tool for the development of decision-support systems that aid medical practitioners in confirming their diagnoses. This method makes it easier for medical professionals to diagnose and initiate prompt treatment while performing preliminary auscultation and reduces unnecessary echocardiograms.

Markov-Based Neural Networks for Heart Sound Segmentation: Using Domain Knowledge in a Principled Way

This work considers the problem of segmenting heart sounds into their fundamental components. We unify statistical and data-driven solutions by introducing Markov-based Neural Networks (MNNs), a hybrid end-to-end framework that exploits Markov models as statistical inductive biases for an Artificial Neural Network (ANN) discriminator. We show that an MNN leveraging a simple one-dimensional Convolutional ANN significantly outperforms two recent purely data-driven solutions for this task in two publicly available datasets: PhysioNet 2016 (Sensitivity: 0.947 ±0.02; Positive Predictive Value : 0.937 ±0.025) and the CirCor DigiScope 2022 (Sensitivity: 0.950 ±0.008; Positive Predictive Value: 0.943 ±0.012). We also propose a novel gradient-based unsupervised learning algorithm that effectively makes the MNN adaptive to unseen datum sampled from unknown distributions. We perform a cross dataset analysis and show that an MNN pre-trained in the CirCor DigiScope 2022 can benefit from an average improvement of 3.90% Positive Predictive Value on unseen observations from the PhysioNet 2016 dataset using this method.

On the Impact of Synchronous Electrocardiogram Signals for Heart Sounds Segmentation

In this paper we study the heart sound segmentation problem using Deep Neural Networks. The impact of available electrocardiogram (ECG) signals in addition to phonocardiogram (PCG) signals is evaluated. To incorporate ECG, two different models considered, which are built upon a 1D U-net - an early fusion one that fuses ECG in an early processing stage, and a late fusion one that averages the probabilities obtained by two networks applied independently on PCG and ECG data. Results show that, in contrast with traditional uses of ECG for PCG gating, early fusion of PCG and ECG information can provide more robust heart sound segmentation. As a proof of concept, we use the publicly available PhysioNet dataset. Validation results provide, on average, a sensitivity of 97.2%, 94.5%, and 95.6% and a Positive Predictive Value of 97.5%, 96.2%, and 96.1% for Early-fusion, Late-fusion, and unimodal (PCG only) models, respectively, showing the advantages of combining both signals at early stages to segment heart sounds.Clinical relevance- Cardiac auscultation is the first line of screening for cardiovascular diseases. Its low cost and simplicity are especially suitable for screening large populations in underprivileged countries. The proposed analysis and algorithm show the potential of effectively including electrocardiogram information to improve heart sound segmentation performance, thus enhancing the capacity of extracting useful information from heart sound recordings.

A novel heart sound segmentation algorithm via multi-feature input and neural network with attention mechanism

Objective. Heart sound segmentation (HSS), which aims to identify the exact positions of the first heart sound(S1), second heart sound(S2), the duration of S1, systole, S2, and diastole within a cardiac cycle of phonocardiogram (PCG), is an indispensable step to find out heart health. Recently, some neural network-based methods for heart sound segmentation have shown good performance.Approach. In this paper, a novel method was proposed for HSS exactly using One-Dimensional Convolution and Bidirectional Long-Short Term Memory neural network with Attention mechanism (C-LSTM-A) by incorporating the 0.5-order smooth Shannon entropy envelope and its instantaneous phase waveform (IPW), and third intrinsic mode function (IMF-3) of PCG signal to reduce the difficulty of neural network learning features.Main results. An average F1-score of 96.85 was achieved in the clinical research dataset (Fuwai Yunnan Cardiovascular Hospital heart sound dataset) and an average F1-score of 95.68 was achieved in 2016 PhysioNet/CinC Challenge dataset using the novel method.Significance. The experimental results show that this method has advantages for normal PCG signals and common pathological PCG signals, and the segmented fundamental heart sound(S1, S2), systole, and diastole signal components are beneficial to the study of subsequent heart sound classification.

Identification of Characteristic Points in Multivariate Physiological Signals by Sensor Fusion and Multi-Task Deep Networks

Identification of characteristic points in physiological signals, such as the peak of the R wave in the electrocardiogram and the peak of the systolic wave of the photopletismogram, is a fundamental step for the quantification of clinical parameters, such as the pulse transit time. In this work, we presented a novel neural architecture, called eMTUnet, to automate point identification in multivariate signals acquired with a chest-worn device. The eMTUnet consists of a single deep network capable of performing three tasks simultaneously: (i) localization in time of characteristic points (labeling task), (ii) evaluation of the quality of signals (classification task); (iii) estimation of the reliability of classification (reliability task). Preliminary results in overnight monitoring showcased the ability to detect characteristic points in the four signals with a recall index of about 1.00, 0.90, 0.90, and 0.80, respectively. The accuracy of the signal quality classification was about 0.90, on average over four different classes. The average confidence of the correctly classified signals, against the misclassifications, was 0.93 vs. 0.52, proving the worthiness of the confidence index, which may better qualify the point identification. From the achieved outcomes, we point out that high-quality segmentation and classification are both ensured, which brings the use of a multi-modal framework, composed of wearable sensors and artificial intelligence, incrementally closer to clinical translation.

Hybrid Convolutional Networks for End-to-End Event Detection in Concurrent PPG and PCG Signals Affected by Motion Artifacts

The accurate detection of physiologically-related events in photopletismographic (PPG) and phocardiographic (PCG) signals, recorded by wearable sensors, is mandatory to perform the estimation of relevant cardiovascular parameters like the heart rate and the blood pressure. However, the measurement performed in uncontrolled conditions without clinical supervision leaves the detection quality particularly susceptible to noise and motion artifacts. The performed work proposed a new fully-automatic computational framework, based on convolutional networks, to identify and localize fiducial points in time as the foot, maximum slope and peak in PPG signal and the S1 sound in the PCG signal, both acquired by a custom chest sensor, described recently in the literature by our group. The novelty entailing a custom neural architecture to process sequentially the PPG and PCG signals. Tests were performed analysing four different acquisition conditions (rest, cycling, rest recovery and walking). Cross-validation results for the three PPG fiducial points showed identification accuracy greater than 93 % and localization error (RMSE) less than 10 ms. As expected, cycling and walking conditions provided worse results than rest and recovery, however reaching an accuracy greater than 90 % and a localization error lower than 15 ms. Likewise, the identification and localization error for S1 sound were greater than 90 % and lower than 25 ms. Overall, this study showcased the ability of the proposed technique to detect events with high accuracy not only for steady acquisitions but also during subject movements. We also showed that the proposed network outperformed traditional Shannon-energy-envelope method in the detection of S1 sound. Therefore, we argue that coupling chest sensors and deep learning processing techniques may disclose wearable devices to unobtrusively acquire health information, being less affected by noise and motion artifacts.

A novel intelligent system based on adjustable classifier models for diagnosing heart sounds

A novel intelligent diagnostic system is proposed to diagnose heart sounds (HSs). The innovations of this system are primarily reflected in the automatic segmentation and extraction of the first complex sound \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$({ CS }_{1})$$\end{document}(CS1) and second complex sound \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$({ CS }_{2})$$\end{document}(CS2); the automatic extraction of the secondary envelope-based diagnostic features \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma _{_1}$$\end{document}γ1, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma _{_2}$$\end{document}γ2, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma _{_3}$$\end{document}γ3 from \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${ CS }_{1}$$\end{document}CS1 and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${ CS }_{2}$$\end{document}CS2; and the adjustable classifier models that correspond to the confidence bounds of the Chi-square (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi ^{2}$$\end{document}χ2) distribution and are adjusted by the given confidence levels (denoted as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\beta$$\end{document}β). The three stages of the proposed system are summarized as follows. In stage 1, the short time modified Hilbert transform (STMHT)-based curve is used to segment and extract \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${ CS }_{1}$$\end{document}CS1 and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${ CS }_{2}$$\end{document}CS2. In stage 2, the envelopes \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${ CS _{1}}_{\mathrm{F_{E}}}$$\end{document}CS1FE and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${ CS _{2}}_{\mathrm{F_{E}}}$$\end{document}CS2FE for periods \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${ CS }_{1}$$\end{document}CS1 and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${ CS }_{2}$$\end{document}CS2 are obtained via a novel method, and the frequency features are automatically extracted from \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${ CS _{1}}_{\mathrm{F_{E}}}$$\end{document}CS1FE and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${ CS _{2}}_{\mathrm{F_{E}}}$$\end{document}CS2FE by setting different threshold value (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$Thv$$\end{document}Thv) lines. Finally, the first three principal components determined based on principal component analysis (PCA) are used as the diagnostic features. In stage 3, a Gaussian mixture model (GMM)-based component objective function \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$f_{ et }(\mathbf{x })$$\end{document}fet(x) is generated. Then, the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi ^{2}$$\end{document}χ2 distribution for component k is determined by calculating the Mahalanobis distance from \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathbf{x }}$$\end{document}x to the class mean \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mu _{_k}$$\end{document}μk for component k, and the confidence region of component k is determined by adjusting the optimal confidence level \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\beta _{k}$$\end{document}βk and used as the criterion to diagnose HSs. The performance evaluation was validated by sounds from online HS databases and clinical heart databases. The accuracy of the proposed method was compared to the accuracies of other state-of-the-art methods, and the highest classification accuracies of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$99.43\%$$\end{document}99.43%, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$98.93\%$$\end{document}98.93%, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$99.13\%$$\end{document}99.13%, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$99.85\%$$\end{document}99.85%, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$98.62\%$$\end{document}98.62%, 99.67\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\%$$\end{document}% and 99.91\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\%$$\end{document}% in the detection of MR, MS, ASD, NM, AS, AR and VSD sounds were achieved by setting \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\beta _{k}(k=1, 2, \ldots , 7)$$\end{document}βk(k=1,2,…,7) to 0.87,0.65,0.67,0.65,0.67,0.79 and 0.87, respectively.

FPGA-Based High-Performance Phonocardiography System for Extraction of Cardiac Sound Components Using Inverse Delayed Neuron Model

The study focuses on the extraction of cardiac sound components using a multi-channel micro-electromechanical system (MEMS) microphone-based phonocardiography system. The proposed multi-channel phonocardiography system classifies the cardiac sound components using artificial neural networks (ANNs) and synaptic weights that are calculated using the inverse delayed (ID) function model of the neuron. The proposed ANN model was simulated in MATLAB^R and implemented in a field-programmable gate array (FPGA). The proposed system examined both abnormal and normal samples collected from 30 patients. Experimental results revealed a good sensitivity of 99.1% and an accuracy of 0.9.

Unsupervised Heart Sound Decomposition and State Estimation with Generative Oscillation Models

This paper proposes a new generative probabilistic model for phonocardiograms (PCGs) that can simultaneously capture oscillatory factors and state transitions in cardiac cycles. Conventionally, PCGs have been modeled in two main aspects. One is a state space model that represents recurrent and frequently appearing state transitions. Another is a factor model that expresses the PCG as a non-stationary signal consisting of multiple oscillations. To model these perspectives in a unified framework, we combine an oscillation decomposition with a state space model. The proposed model can decompose the PCG into cardiac state dependent oscillations by reflecting the mechanism of cardiac sounds generation in an unsupervised manner. In the experiments, our model achieved better accuracy in the state estimation task compared to the empirical mode decomposition method. In addition, our model detected S2 onsets more accurately than the supervised segmentation method when distributions among PCG signals were different.

Cardiovascular Disease Recognition Based on Heartbeat Segmentation and Selection Process

Assessment of heart sounds which are generated by the beating heart and the resultant blood flow through it provides a valuable tool for cardiovascular disease (CVD) diagnostics. The cardiac auscultation using the classical stethoscope phonological cardiogram is known as the most famous exam method to detect heart anomalies. This exam requires a qualified cardiologist, who relies on the cardiac cycle vibration sound (heart muscle contractions and valves closure) to detect abnormalities in the heart during the pumping action. Phonocardiogram (PCG) signal represents the recording of sounds and murmurs resulting from the heart auscultation, typically with a stethoscope, as a part of medical diagnosis. For the sake of helping physicians in a clinical environment, a range of artificial intelligence methods was proposed to automatically analyze PCG signal to help in the preliminary diagnosis of different heart diseases. The aim of this research paper is providing an accurate CVD recognition model based on unsupervised and supervised machine learning methods relayed on convolutional neural network (CNN). The proposed approach is evaluated on heart sound signals from the well-known, publicly available PASCAL and PhysioNet datasets. Experimental results show that the heart cycle segmentation and segment selection processes have a direct impact on the validation accuracy, sensitivity (TPR), precision (PPV), and specificity (TNR). Based on PASCAL dataset, we obtained encouraging classification results with overall accuracy 0.87, overall precision 0.81, and overall sensitivity 0.83. Concerning Micro classification results, we obtained Micro accuracy 0.91, Micro sensitivity 0.83, Micro precision 0.84, and Micro specificity 0.92. Using PhysioNet dataset, we achieved very good results: 0.97 accuracy, 0.946 sensitivity, 0.944 precision, and 0.946 specificity.

Domain Generalization in Biosignal Classification

OBJECTIVE

When training machine learning models, we often assume that the training data and evaluation data are sampled from the same distribution. However, this assumption is violated when the model is evaluated on another unseen but similar database, even if that database contains the same classes. This problem is caused by domain-shift and can be solved using two approaches: domain adaptation and domain generalization. Simply, domain adaptation methods can access data from unseen domains during training; whereas in domain generalization, the unseen data is not available during training. Hence, domain generalization concerns models that perform well on inaccessible, domain-shifted data.

METHOD

Our proposed domain generalization method represents an unseen domain using a set of known basis domains, afterwhich we classify the unseen domain using classifier fusion. To demonstrate our system, we employ a collection of heart sound databases that contain normal and abnormal sounds (classes).

RESULTS

Our proposed classifier fusion method achieves accuracy gains of up to 16% for four completely unseen domains.

CONCLUSION

Recognizing the complexity induced by the inherent temporal nature of biosignal data, the two-stage method proposed in this study is able to effectively simplify the whole process of domain generalization while demonstrating good results on unseen domains and the adopted basis domains.

SIGNIFICANCE

To our best knowledge, this is the first study that investigates domain generalization for biosignal data. Our proposed learning strategy can be used to effectively learn domain-relevant features while being aware of the class differences in the data.

End-to-end heart sound segmentation using deep convolutional recurrent network

Heart sound segmentation (HSS) aims to detect the four stages (first sound, systole, second heart sound and diastole) from a heart cycle in a phonocardiogram (PCG), which is an essential step in automatic auscultation analysis. Traditional HSS methods need to manually extract the features before dealing with HSS tasks. These artificial features highly rely on extraction algorithms, which often result in poor performance due to the different operating environments. In addition, the high-dimension and frequency characteristics of audio also challenge the traditional methods in effectively addressing HSS tasks. This paper presents a novel end-to-end method based on convolutional long short-term memory (CLSTM), which directly uses audio recording as input to address HSS tasks. Particularly, the convolutional layers are designed to extract the meaningful features and perform the downsampling, and the LSTM layers are developed to conduct the sequence recognition. Both components collectively improve the robustness and adaptability in processing the HSS tasks. Furthermore, the proposed CLSTM algorithm is easily extended to other complex heart sound annotation tasks, as it does not need to extract the characteristics of corresponding tasks in advance. In addition, the proposed algorithm can also be regarded as a powerful feature extraction tool, which can be integrated into the existing models for HSS. Experimental results on real-world PCG datasets, through comparisons to peer competitors, demonstrate the outstanding performance of the proposed algorithm.

Estimation of heart rate from tracheal sounds recorded for the sleep apnea syndrome diagnosis

Obstructive sleep apnea is a common sleep disorder with a high prevalence and often accompanied by significant snoring activity. To diagnose this condition, polysomnography is the standard method, where a neck microphone could be added to record tracheal sounds. These can then be used to study the characteristics of breathing, snoring or apnea. In addition cardiac sounds, also present in the acquired data, could be exploited to extract heart rate. The paper presents new algorithms for estimating heart rate from tracheal sounds, especially in very loud snoring environment. The advantage is that it is possible to reduce the number of diagnostic devices, especially for compact home applications. Three algorithms are proposed, based on optimal filtering and cross-correlation. They are tested firstly on one patient presenting significant pathology of apnea syndrome, with a recording of 509 min. Secondly, an extension to a database of 16 patients is proposed (16 hours of recording). When compared to a reference ECG signal, the final results obtained from tracheal sounds reach an accuracy of 81% to 98% and an RMS error from 1.3 to 4.2 bpm, according to the level of snoring and to the considered algorithm.

Research on Segmentation and Classification of Heart Sound Signals Based on Deep Learning

The heart sound signal is one of the signals that reflect the health of the heart. Research on the heart sound signal contributes to the early diagnosis and prevention of cardiovascular diseases. As a commonly used deep learning network, convolutional neural network (CNN) has been widely used in images. In this paper, the method of analyzing heart sound through using CNN has been studied. Firstly, the original data set was preprocessed, and then the heart sounds were segmented on U-net, based on the deep CNN. Finally, the classification of heart sounds was completed through CNN. The data from 2016 PhysioNet/CinC Challenge was utilized for algorithm validation, and the following results were obtained. When the heart sound segmented, the overall accuracy rate was 0.991, the accuracy of the first heart sound was 0.991, the accuracy of the systolic period was 0.996, the accuracy of the second heart sound was 0.996, and the accuracy of the diastolic period was 0.997, and the average accuracy rate was 0.995; While in classification, the accuracy was 0.964, the sensitivity was 0.781, and the specificity was 0.873. These results show that deep learning based on CNN shows good performance in the segmentation and classification of the heart sound signal.

Temporal Convolutional Network Connected with an Anti-Arrhythmia Hidden Semi-Markov Model for Heart Sound Segmentation

Heart sound segmentation (HSS) is a critical step in heart sound processing, where it improves the interpretability of heart sound disease classification algorithms. In this study, we aimed to develop a real-time algorithm for HSS by combining the temporal convolutional network (TCN) and the hidden semi-Markov model (HSMM), and improve the performance of HSMM for heart sounds with arrhythmias. We experimented with TCN and determined the best parameters based on spectral features, envelopes, and one-dimensional CNN. However, the TCN results could contradict the natural fixed order of S1-systolic-S2-diastolic of heart sound, and thereby the Viterbi algorithm based on HSMM was connected to correct the order errors. On this basis, we improved the performance of the Viterbi algorithm when detecting heart sounds with cardiac arrhythmias by changing the distribution and weights of the state duration probabilities. The public PhysioNet Computing in Cardiology Challenge 2016 data set was employed to evaluate the performance of the proposed algorithm. The proposed algorithm achieved an F1 score of 97.02%, and this result was comparable with the current state-of-the-art segmentation algorithms. In addition, the proposed enhanced Viterbi algorithm for HSMM corrected 30 out of 30 arrhythmia errors after checking one by one in the dataset.

Evaluation of Hemodialysis Arteriovenous Bruit by Deep Learning

Physical findings of auscultation cannot be quantified at the arteriovenous fistula examination site during daily dialysis treatment. Consequently, minute changes over time cannot be recorded based only on subjective observations. In this study, we sought to supplement the daily arteriovenous fistula consultation for hemodialysis patients by recording the sounds made by the arteriovenous fistula and evaluating the sounds using deep learning methods to provide an objective index. We sampled arteriovenous fistula auscultation sounds (192 kHz, 24 bits) recorded over 1 min from 20 patients. We also extracted arteriovenous fistula sounds for each heartbeat without environmental sound by using a convolutional neural network (CNN) model, which was made by comparing these sound patterns with 5000 environmental sounds. The extracted single-heartbeat arteriovenous fistula sounds were sent to a spectrogram and scored using a CNN learning model with bidirectional long short-term memory, in which the degree of arteriovenous fistula stenosis was assigned to one of five sound types (i.e., normal, hard, high, intermittent, and whistling). After 100 training epochs, the method exhibited an accuracy rate of 70-93%. According to the receiver operating characteristic (ROC) curve, the area under the ROC curves (AUC) was 0.75-0.92. The analysis of arteriovenous fistula sound using deep learning has the potential to be used as an objective index in daily medical care.

Temporal-Framing Adaptive Network for Heart Sound Segmentation Without Prior Knowledge of State Duration

OBJECTIVE

This paper presents a novel heart sound segmentation algorithm based on Temporal-Framing Adaptive Network (TFAN), including state transition loss and dynamic inference.

METHODS

In contrast to previous state-of-the-art approaches, TFAN does not require any prior knowledge of the state duration of heart sounds and is therefore likely to generalize to non sinus rhythm. TFAN was trained on 50 recordings randomly chosen from Training set A of the 2016 PhysioNet/Computer in Cardiology Challenge and tested on the other 12 independent databases (2,099 recordings and 52,180 beats). And further testing of performance was conducted on databases with three levels of increasing difficulty (LEVEL-I, -II and -III).

RESULTS

TFAN achieved a superior F₁ score for all 12 databases except for 'Test-B,' with an average of 96.72%, compared to 94.56% for logistic regression hidden semi-Markov model (LR-HSMM) and 94.18% for bidirectional gated recurrent neural network (BiGRNN). Moreover, TFAN achieved an overall F₁ score of 99.21%, 94.17%, 91.31% on LEVEL-I, -II and -III databases respectively, compared to 98.37%, 87.56%, 78.46% for LR-HSMM and 99.01%, 92.63%, 88.45% for BiGRNN.

CONCLUSION

TFAN therefore provides a substantial improvement on heart sound segmentation while using less parameters compared to BiGRNN.

SIGNIFICANCE

The proposed method is highly flexible and likely to apply to other non-stationary time series. Further work is required to understand to what extent this approach will provide improved diagnostic performance, although it is logical to assume superior segmentation will lead to improved diagnostics.

Multi-channel lung sound classification with convolutional recurrent neural networks

In this paper, we present an approach for multi-channel lung sound classification, exploiting spectral, temporal and spatial information. In particular, we propose a frame-wise classification framework to process full breathing cycles of multi-channel lung sound recordings with a convolutional recurrent neural network. With our recently developed 16-channel lung sound recording device, we collect lung sound recordings from lung-healthy subjects and patients with idiopathic pulmonary fibrosis (IPF), within a clinical trial. From the lung sound recordings, we extract spectrogram features and compare different deep neural network architectures for binary classification, i.e. healthy vs. pathological. Our proposed classification framework with the convolutional recurrent neural network outperforms the other networks by achieving an F-score of F₁≈92%. Together with our multi-channel lung sound recording device, we present a holistic approach to multi-channel lung sound analysis.

Enhancement of the heart sound envelope using the logistic function amplitude moderation method

This paper presents a new method to extract the envelope of the fundamental heart sound (S1 and S2) using the logistic function. The sigmoid characteristic of the logistic function is incorporated to segregate S1, and S2 signal intensities from silent or noise interfered systolic and diastolic intervals in a heart sound cycle. This signal intensity transformation brings uniformity to the envelope peak of S1 and S2 sound by inclining the transform intensity distribution towards the upper asymptote of the sigmoid curve. The proposed logistic function based amplitude moderation (LFAM) envelogram method involves finding the critical upper amplitude (x_uc) above which the signals will be categorized as loud sound and the critical lower amplitude (x_lc) below which the signal will be considered as noise. These critical values are regressively obtained from the signal itself by histogram analysis of intensity distribution. The performance is evaluated on noisy PCG dataset taken from PhysioNet/Computing in Cardiology Challenge 2016. The LFAM envelope yields better hill-valley discrimination of heart sounds from its silent/noisy signal intervals. The enhance heart sound envelope peaks are better than conventional methods. The proposed envelope feature is evaluated for heart sound segmentation using HSMM. There is a significant improvement in segmentation accuracy, especially at a low signal-to-noise ratio. The best average F1 score is 97.73%.

Towards Domain Invariant Heart Sound Abnormality Detection Using Learnable Filterbanks

OBJECTIVE

Cardiac auscultation is the most practiced non-invasive and cost-effective procedure for the early diagnosis of heart diseases. While machine learning based systems can aid in automatically screening patients, the robustness of these systems is affected by numerous factors including the stethoscope/sensor, environment, and data collection protocol. This article studies the adverse effect of domain variability on heart sound abnormality detection and develops strategies to address this problem.

METHODS

We propose a novel Convolutional Neural Network (CNN) layer, consisting of time-convolutional (tConv) units, that emulate Finite Impulse Response (FIR) filters. The filter coefficients can be updated via backpropagation and be stacked in the front-end of the network as a learnable filterbank.

RESULTS

On publicly available multi-domain datasets, the proposed method surpasses the top-scoring systems found in the literature for heart sound abnormality detection (a binary classification task). We utilized sensitivity, specificity, F-1 score and Macc (average of sensitivity and specificity) as performance metrics. Our systems achieved relative improvements of up to 11.84% in terms of MAcc, compared to state-of-the-art methods.

CONCLUSION

The results demonstrate the effectiveness of the proposed learnable filterbank CNN architecture in achieving robustness towards sensor/domain variability in PCG signals.

SIGNIFICANCE

The proposed methods pave the way for deploying automated cardiac screening systems in diversified and underserved communities.

Segmentation of Radar-Recorded Heart Sound Signals Using Bidirectional LSTM Networks

Sounds caused by the action of the heart reflect both its health as well as deficiencies and are examined by physicians since antiquity. Pathologies of the valves, e.g. insufficiencies and stenosis, cardiac effusion, arrhythmia, inflammation of the surrounding tissue and other diagnosis can be reached by experienced physicians. However, practice is needed to assess the findings correctly. Furthermore, stethoscopes do not allow for long-term monitoring of a patient. Recently, radar technology has shown the ability to perform continuous touchless and thereby burden-free heart sound measurements. In order to perform automated classification of the signals, the first and most important step is to segment the heart sounds into their physiological phases. This paper examines the use of different Long Short-Term Memory (LSTM) architectures for this purpose based on a large dataset of radar-recorded heart sounds gathered from 30 different test persons in a clinical study. The best-performing network, a bidirectional LSTM, achieves a sample-wise accuracy of 93.4 % and a F1 score for the first heart sound of 95.8 %.

A Review of Computer-Aided Heart Sound Detection Techniques

Cardiovascular diseases have become one of the most prevalent threats to human health throughout the world. As a noninvasive assistant diagnostic tool, the heart sound detection techniques play an important role in the prediction of cardiovascular diseases. In this paper, the latest development of the computer-aided heart sound detection techniques over the last five years has been reviewed. There are mainly the following aspects: the theories of heart sounds and the relationship between heart sounds and cardiovascular diseases; the key technologies used in the processing and analysis of heart sound signals, including denoising, segmentation, feature extraction and classification; with emphasis, the applications of deep learning algorithm in heart sound processing. In the end, some areas for future research in computer-aided heart sound detection techniques are explored, hoping to provide reference to the prediction of cardiovascular diseases.

Machine Listening for Heart Status Monitoring: Introducing and Benchmarking HSS - the Heart Sounds Shenzhen Corpus

Auscultation of the heart is a widely studied technique, which requires precise hearing from practitioners as a means of distinguishing subtle differences in heart-beat rhythm. This technique is popular due to its non-invasive nature, and can be an early diagnosis aid for a range of cardiac conditions. Machine listening approaches can support this process, monitoring continuously and allowing for a representation of both mild and chronic heart conditions. Despite this potential, relevant databases and benchmark studies are scarce. In this paper, we introduce our publicly accessible database, the Heart Sounds Shenzhen Corpus (HSS), which was first released during the recent INTERSPEECH 2018 ComParE Heart Sound sub-challenge. Additionally, we provide a survey of machine learning work in the area of heart sound recognition, as well as a benchmark for HSS utilising standard acoustic features and machine learning models. At best our support vector machine with Log Mel features achieves 49.7% unweighted average recall on a three category task (normal, mild, moderate/severe).

Heart Sound Segmentation Using Bidirectional LSTMs With Attention

OBJECTIVE

This paper proposes a novel framework for the segmentation of phonocardiogram (PCG) signals into heart states, exploiting the temporal evolution of the PCG as well as considering the salient information that it provides for the detection of the heart state.

METHODS

We propose the use of recurrent neural networks and exploit recent advancements in attention based learning to segment the PCG signal. This allows the network to identify the most salient aspects of the signal and disregard uninformative information.

RESULTS

The proposed method attains state-of-the-art performance on multiple benchmarks including both human and animal heart recordings. Furthermore, we empirically analyse different feature combinations including envelop features, wavelet and Mel Frequency Cepstral Coefficients (MFCC), and provide quantitative measurements that explore the importance of different features in the proposed approach.

CONCLUSION

We demonstrate that a recurrent neural network coupled with attention mechanisms can effectively learn from irregular and noisy PCG recordings. Our analysis of different feature combinations shows that MFCC features and their derivatives offer the best performance compared to classical wavelet and envelop features.

SIGNIFICANCE

Heart sound segmentation is a crucial pre-processing step for many diagnostic applications. The proposed method provides a cost effective alternative to labour extensive manual segmentation, and provides a more accurate segmentation than existing methods. As such, it can improve the performance of further analysis including the detection of murmurs and ejection clicks. The proposed method is also applicable for detection and segmentation of other one dimensional biomedical signals.

CLASSIFICATION OF UNSEGMENTED HEART SOUND RECORDING USING KNN CLASSIFIER

Due to low physical workout, high-calorie intake, and bad behavioral character, people were affected by cardiological disorders. Every instant, one out of four deaths are due to heart-related ailments. Hence, the early diagnosis of a heart is essential. Most of the approaches for automated classification of the heart sound need segmentation of Phonocardiograms (PCG) signal. The main aim of this study was to decline the segmentation process and to estimate the utility for accurate and detailed classification of short unsegmented PCG recording. Based on wavelet decomposition, Hilbert transform, homomorphic filtering, and power spectral density (PSD), the features had been obtained using the beginning 5 second PCG recording. The extracted features were classified using nearest neighbors with Euclidean distances for different values of [Formula: see text] by bootstrapping 50% PCG recording for training and 50% for testing over 100 iterations. The overall accuracy of 100%, 85%, 80.95%, 81.4%, and 98.13% had been achieved for five different datasets using KNN classifiers. The classification performance for analyzing the whole datasets is 90% accuracy with 93% sensitivity and 90% specificity. The classification of unsegmented PCG recording based on an efficient feature extraction is necessary. This paper presents a promising classification performance as compared with the state-of-the-art approaches in short time less complexity.