Heart Sound Recordings Research Articles

Phonocardiogram (PCG) Murmur Detection Based on the Mean Teacher Method.

Cardiovascular diseases (CVDs) are among the primary causes of mortality globally, highlighting the critical need for early detection to mitigate their impact. Phonocardiograms (PCGs), which record heart sounds, are essential for the non-invasive assessment of cardiac function, enabling the early identification of abnormalities such as murmurs. Particularly in underprivileged regions with high birth rates, the absence of early diagnosis poses a significant public health challenge. In pediatric populations, the analysis of PCG signals is invaluable for detecting abnormal sound waves indicative of congenital and acquired heart diseases, such as septal defects and defective cardiac valves. In the PhysioNet 2022 challenge, the murmur score is a weighted accuracy metric that reflects detection accuracy based on clinical significance. In our research, we proposed a mean teacher method tailored for murmur detection, making full use of the Phyionet2022 and Phyionet2016 PCG datasets, achieving the SOTA (State of Art) performance with a murmur score of 0.82 and an AUC score of 0.90, providing an accessible and high accuracy non-invasive early stage CVD assessment tool, especially for low and middle-income countries (LMICs).

Evaluating heart rate variability by a novel stethoscopic approach to minimise stress in cats.

Stress associated with manipulation during electrocardiography (ECG) recording in cats potentially limits the assessment of autonomic function through heart rate variability (HRV) in the feline population. This study proposed an alternative, cat friendly, stethoscopic approach to evaluate HRV with an easily acquired vasovagal tonus index (VVTI). The aim of this prospective study was to evaluate whether VVTI derived from heart sound signals could distinguish between relaxed and stimulated states. A total of 29 cats with 56 recordings of heart sound and ECG on 31 occasions were included. In 25 cats in their home environment, a stethoscope connected to a digital recording device was used to record 2 mins of heart sounds twice - with the cats in a relaxed state and immediately after stimulation. The VVTI was calculated from 20, 60 and 120 consecutive beat-to-beat intervals on the heart sound spectrogram (stethoscopic-VVTI 20, 60 and 120), using the natural logarithm of the variance of the intervals based on previous literature. A 2-min ECG recording was obtained at home with the intention of avoiding strict restraint. To demonstrate the feasibility of the stethoscopic approach in a hospital setting, six cats (two of which were also recorded at home) underwent heart sound and ECG recordings during planned veterinary visits. Stethoscopic-VVTI 20 (5.43 to 4.79, P = 0.001), 60 (6.20 to 5.18, P <0.001) and 120 (6.24 to 5.60, P = 0.02) all significantly decreased after stimulation, indicating a reduced vasovagal tone as expected. Calculations of stethoscopic-VVTI from different sections of the recording yielded statistically similar results. Stethoscopic-VVTI showed a negative correlation with the corresponding heart rate. Bland-Altman analysis revealed a mean bias for the differences between stethoscopic-VVTI and ECG-VVTI of 0.50 and 1.07 at home and in the hospital, respectively. VVTI can be successfully detected through a stethoscopic approach, serving as a less stressful tool for HRV evaluation in cats during routine auscultation.

PCGmix: A Data-Augmentation Method For Heart-Sound Classification.

Cardiovascular diseases (CVDs) are a leading cause of mortality worldwide, responsible for 32% of all deaths, with the annual death toll projected to reach 23.3 million by 2030. The early identification of individuals at high risk of CVD is crucial for the effectiveness of preventive strategies. In the field of deep learning, automated CVD-detection methods have gained traction, with phonocardiogram (PCG) data emerging as a valuable resource. However, deep-learning models rely on large datasets, which are often challenging to obtain. In recent years, data augmentation has become a viable solution to the problem of scarce data. In this paper, we propose a novel data-augmentation technique named PCGmix, specifically engineered for the augmentation of PCG data. The PCGmix algorithm employs a process of segmenting and reassembling PCG recordings, incorporating meticulous interpolation to ensure the preservation of the cardinal diagnostic features pertinent to CVD detection. The empirical assessment of the PCGmix method was utilized on a publicly available database of normal and abnormal heart-sound recordings. To evaluate the impact of data augmentation across a range of dataset sizes, we conducted experiments encompassing both limited and extensive amounts of training data. The experimental results demonstrate that the novel method is superior to the compared state-of-the-art, time-series augmentation. Notably, on limited data, our method achieves comparable accuracy to the no-augmentation approach when trained on 31% to 69% larger datasets. This study suggests that PCGmix can enhance the accuracy of deep-learning models for CVD detection, especially in data-constrained environments.

A Robust Deep Learning Framework Based on Spectrograms for Heart Sound Classification.

Heart sound analysis plays an important role in early detecting heart disease. However, manual detection requires doctors with extensive clinical experience, which increases uncertainty for the task, especially in medically underdeveloped areas. This paper proposes a robust neural network structure with an improved attention module for automatic classification of heart sound wave. In the preprocessing stage, noise removal with Butterworth bandpass filter is first adopted, and then heart sound recordings are converted into time-frequency spectrum by short-time Fourier transform (STFT). The model is driven by STFT spectrum. It automatically extracts features through four down sample blocks with different filters. Subsequently, an improved attention module based on Squeeze-and-Excitation module and coordinate attention module is developed for feature fusion. Finally, the neural network will give a category for heart sound waves based on the learned features. The global average pooling layer is adopted for reducing the model's weight and avoiding overfitting, while focal loss is further introduced as the loss function to minimize the data imbalance problem. Validation experiments have been conducted on two publicly available datasets, and the results well demonstrate the effectiveness and advantages of our method.

Evaluation of a digitally enhanced cardiac auscultation learning method in Cameroon: results of a controlled study

SummaryBackgroundCardiac auscultation is an efficient and effective diagnostic tool, especially in low-income countries where access to modern diagnostic methods remains difficult. This study aimed to evaluate the effect of a digitally enhanced cardiac auscultation learning method on medical students’ performance and satisfaction.MethodsWe conducted a double-arm parallel controlled trial, including newly admitted 4th -year medical students enrolled in two medical schools in Yaoundé, Cameroon and allocated into two groups: the intervention group (benefiting from theoretical lessons, clinical internship and the listening sessions of audio recordings of heart sounds) and the control group (benefiting from theoretical lessons and clinical internship). All the participants were subjected to a pretest before the beginning of the training, evaluating theoretical knowledge and recognition of cardiac sounds, and a post-test at the eighth week of clinical training associated with the evaluation of satisfaction. The endpoints were the progression of knowledge score, skills score, total (knowledge and skills) score and participant satisfaction.ResultsForty-nine participants (27 in the intervention group and 22 in the control group) completed the study. The knowledge progression (+ 26.7 versus + 7.5; p ˂0.01) and the total progression (+ 22.5 versus + 14.6; p ˂ 0.01) were higher in the intervention group with a statistically significant difference compared to the control group. There was no significant difference between the two groups regarding skills progression (+ 25 versus + 17.5; p = 0.27). Satisfaction was higher in general in the intervention group (p ˂ 0.01), which recommended this method compared to the control group.ConclusionThe learning method of cardiac auscultation reinforced by the listening sessions of audio recordings of heart sounds improves medical students’ performances (knowledge and global – knowledge and skills) who find it satisfactory and recommendable.Trial RegistrationThis trial has been registered the 29/11/2019 in the Pan African Clinical Trials Registry (http://www.pactr.org) under unique identification number PACTR202001504666847 and the protocol has been published in BMC Medical Education.

Acquisition Devices for Fetal Phonocardiography: A Scoping Review.

Timely and reliable fetal monitoring is crucial to prevent adverse events during pregnancy and delivery. Fetal phonocardiography, i.e., the recording of fetal heart sounds, is emerging as a novel possibility to monitor fetal health status. Indeed, due to its passive nature and its noninvasiveness, the technique is suitable for long-term monitoring and for telemonitoring applications. Despite the high share of literature focusing on signal processing, no previous work has reviewed the technological hardware solutions devoted to the recording of fetal heart sounds. Thus, the aim of this scoping review is to collect information regarding the acquisition devices for fetal phonocardiography (FPCG), focusing on technical specifications and clinical use. Overall, PRISMA-guidelines-based analysis selected 57 studies that described 26 research prototypes and eight commercial devices for FPCG acquisition. Results of our review study reveal that no commercial devices were designed for fetal-specific purposes, that the latest advances involve the use of multiple microphones and sensors, and that no quantitative validation was usually performed. By highlighting the past and future trends and the most relevant innovations from both a technical and clinical perspective, this review will represent a useful reference for the evaluation of different acquisition devices and for the development of new FPCG-based systems for fetal monitoring.

Identifying pediatric heart murmurs and distinguishing innocent from pathologic using deep learning

ObjectiveTo develop a deep learning algorithm to perform multi-class classification of normal pediatric heart sounds, innocent murmurs, and pathologic murmurs. MethodsWe prospectively enrolled children under age 18 being evaluated by the Division of Pediatric Cardiology. Parents provided consent for a deidentified recording of their child's heart sounds with a digital stethoscope. Innocent murmurs were validated by a pediatric cardiologist and pathologic murmurs were validated by echocardiogram. To augment our collection of normal heart sounds, we utilized a public database of pediatric heart sound recordings (Oliveira, 2022). We propose two novel approaches for this audio classification task. We train a vision transformer on either Markov transition field or Gramian angular field image representations of the frequency spectrum. We benchmark our results against a ResNet-50 CNN trained on spectrogram images. ResultsOur final dataset consisted of 366 normal heart sounds, 175 innocent murmurs, and 216 pathologic murmurs. Innocent murmurs collected include Still's murmur, venous hum, and flow murmurs. Pathologic murmurs included ventricular septal defect, tetralogy of Fallot, aortic regurgitation, aortic stenosis, pulmonary stenosis, mitral regurgitation and stenosis, and tricuspid regurgitation. We find that the Vision Transformer consistently outperforms the ResNet-50 on all three image representations, and that the Gramian angular field is the superior image representation for pediatric heart sounds. We calculated a one-vs-rest multi-class ROC curve for each of the three classes. Our best model achieves an area under the curve (AUC) value of 0.92 ± 0.05, 0.83 ± 0.04, and 0.88 ± 0.04 for identifying normal heart sounds, innocent murmurs, and pathologic murmurs, respectively. ConclusionWe present two novel methods for pediatric heart sound classification, which outperforms the current standard of using a convolutional neural network trained on spectrogram images. To our knowledge, we are the first to demonstrate multi-class classification of pediatric murmurs. Multiclass output affords a more explainable and interpretable model, which can facilitate further model improvement in the downstream model development cycle and enhance clinician trust and therefore adoption.

Algorithm for predicting valvular heart disease from heart sounds in an unselected cohort.

This study aims to assess the ability of state-of-the-art machine learning algorithms to detect valvular heart disease (VHD) from digital heart sound recordings in a general population that includes asymptomatic cases and intermediate stages of disease progression. We trained a recurrent neural network to predict murmurs from heart sound audio using annotated recordings collected with digital stethoscopes from four auscultation positions in 2,124 participants from the Tromsø7 study. The predicted murmurs were used to predict VHD as determined by echocardiography. The presence of aortic stenosis (AS) was detected with a sensitivity of 90.9%, a specificity of 94.5%, and an area under the curve (AUC) of 0.979 (CI: 0.963-0.995). At least moderate AS was detected with an AUC of 0.993 (CI: 0.989-0.997). Moderate or greater aortic and mitral regurgitation (AR and MR) were predicted with AUC values of 0.634 (CI: 0.565-703) and 0.549 (CI: 0.506-0.593), respectively, which increased to 0.766 and 0.677 when clinical variables were added as predictors. The AUC for predicting symptomatic cases was higher for AR and MR, 0.756 and 0.711, respectively. Screening jointly for symptomatic regurgitation or presence of stenosis resulted in an AUC of 0.86, with 97.7% of AS cases (n = 44) and all 12 MS cases detected. The algorithm demonstrated excellent performance in detecting AS in a general cohort, surpassing observations from similar studies on selected cohorts. The detection of AR and MR based on HS audio was poor, but accuracy was considerably higher for symptomatic cases, and the inclusion of clinical variables improved the performance of the model significantly.

Learning Representations from Heart Sound: A Comparative Study on Shallow and Deep Models.

Leveraging the power of artificial intelligence to facilitate an automatic analysis and monitoring of heart sounds has increasingly attracted tremendous efforts in the past decade. Nevertheless, lacking on standard open-access database made it difficult to maintain a sustainable and comparable research before the first release of the PhysioNet CinC Challenge Dataset. However, inconsistent standards on data collection, annotation, and partition are still restraining a fair and efficient comparison between different works. To this line, we introduced and benchmarked a first version of the Heart Sounds Shenzhen (HSS) corpus. Motivated and inspired by the previous works based on HSS, we redefined the tasks and make a comprehensive investigation on shallow and deep models in this study. First, we segmented the heart sound recording into shorter recordings (10 s), which makes it more similar to the human auscultation case. Second, we redefined the classification tasks. Besides using the 3 class categories (normal, moderate, and mild/severe) adopted in HSS, we added a binary classification task in this study, i.e., normal and abnormal. In this work, we provided detailed benchmarks based on both the classic machine learning and the state-of-the-art deep learning technologies, which are reproducible by using open-source toolkits. Last but not least, we analyzed the feature contributions of best performance achieved by the benchmark to make the results more convincing and interpretable.

Smartphone for heart sound measurement in hospital: which factors affect heart sound quality?

Abstract Background Smartphone has recently been utilized to measure heart sounds in general population, but not in hospitalized patients. Purpose To explore the factors affecting the quality of heart sounds collected from the patients using smartphone in hospital. Methods An App was developed for Android smartphone to use the built-in microphone for measuring heart sounds on cardiac apex for 20 – 30 seconds in patients. Heart sounds were visually assessed for the presence of the first and second heart sounds and categorized into good, uncertain or bad quality. The factors affecting the percentage of good-quality heart sound recordings were analyzed, including the patients’ baseline characteristics, medical history and blood test. Results Among the 296 patients enrolled, 267 (90%) patients had a good-quality heart sound recording. Good-quality heart sounds were less recorded in emergency room (76%), cardiac care unit (82%), intensive care unit (87%) and neurosurgical intensive care unit (88%). Sex, body mass index, body position (supine vs. sitting), heart rate, medical history of atrial fibrillation, coronary artery disease, heart failure or chronic obstructive pulmonary disease, did not significantly affect heart sound quality (all P &amp;gt; 0.05). An advanced age was related to a lower percentage of good-quality heart sounds (&amp;lt; 60 vs. ≥ 60 years: 94% vs. 87%, P = 0.048). Heart sound quality significantly decreased with red blood cell count, hemoglobin level and hematocrit (all P &amp;lt; 0.001). Conclusions Smartphone can be used in hospital settings for measuring heart sounds in good quality, but factors such as department of hospitalization, age and blood indicators need to be considered.Central illustration

DS-CNN: Dual-Stream Convolutional Neural Networks-Based Heart Sound Classification for Wearable Devices

Cardiovascular diseases (CVDs) is considered a serious public health problem due to the uncertainty of its onset. Consuming wearable devices have increasing popularities for healthcare monitoring, and many of them are capable of continuous monitoring and early detection of CVDs. This paper proposes a framework for heart sound detection that can be considered for deployment on smart wearable devices to screen CVDs conveniently. A dual-stream convolutional neural network (DS-CNN) is developed to detect abnormal ones from short-term heart sound recordings. Preprocessing module is first employed for noise filtering and amplitude normalization. Then short-time Fourier transform and higher-order spectral are introduced for feature extraction, whose products are subsequently fed into the DS-CNN for screening abnormal heart sound signals. Two open accessible datasets are employed for performance evaluation. The results well demonstrate the classification accuracy of the proposed DS-CNN, and also indicate its advantages for adapting to heart sound recordings collected by different equipments.

Real-Time Implementation of a Frequency Shifter for Enhancement of Heart Sounds Perception on VLIW DSP Platform

Auscultation of heart sounds is important to perform cardiovascular assessment. External noises may limit heart sound perception. In addition, heart sound bandwidth is concentrated at very low frequencies, where the human ear has poor sensitivity. Therefore, the acoustic perception of the operator can be significantly improved by shifting the heart sound spectrum toward higher frequencies. This study proposes a real-time frequency shifter based on the Hilbert transform. Key system components are the Hilbert transformer implemented as a Finite Impulse Response (FIR) filter, and a Direct Digital Frequency Synthesizer (DDFS), which allows agile modification of the frequency shift. The frequency shifter has been implemented on a VLIW Digital Signal Processor (DSP) by devising a novel piecewise quadratic approximation technique for efficient DDFS implementation. The performance has been compared with other DDFS implementations both considering piecewise linear technique and sine/cosine standard library functions of the DSP. Piecewise techniques allow a more than 50% reduction in execution time compared to the DSP library. Piecewise quadratic technique also allows a more than 50% reduction in total required memory size in comparison to the piecewise linear. The theoretical analysis of the dynamic power dissipation exhibits a more than 20% reduction using piecewise techniques with respect to the DSP library. The real-time operation has been also verified on the DSK6713 rapid prototyping board by Texas Instruments C6713 DSP. Audiologic tests have also been performed to assess the actual improvement of heart sound perception. To this aim, heart sound recordings were corrupted by additive white Gaussian noise, crowded street noise, and helicopter noise, with different signal-to-noise ratios. All recordings were collected from public databases. Statistical analyses of the audiological test results confirm that the proposed approach provides a clear improvement in heartbeat perception in noisy environments.

Heart Sounds Classification Based on High‐Order Spectrogram and Multi‐Convolutional Neural Network after a New Screening Strategy

AbstractThis paper proposes a pre‐processing method for heart sound screening and extracts the high‐order spectral feature of phonocardiogram. Moreover, a multi‐convolutional neural network (mCNN) is constructed to achieve the classification of normal, aortic stenosis, mitral regurgitation, mitral stenosis, and mitral valve prolapse. First, the heart sound recordings are down‐sampled, denoised by wavelet transform, and normalized. Second, a new heart sound screening algorithm is proposed. The waveform of the heart sound recording is segmented and saved as an image which is performed by the gray‐scale processing to calculate the amplitude of the heart sound. The extremely noisy heart sound segments are screened out based on the amplitude information, and the remaining heart sound segments are spliced as pure heart sound recordings. After 50% superposition segmentation of the heart sound recordings, high‐order spectral features are extracted and image data are stored. Finally, a 34‐layer mCNN is specifically designed to boost the performance of heart sound classification through multi‐layer dimensionality reduction. Experimental results show that the proposed method has superior performance compared with the existing one. For the two‐category dataset, the accuracy with and without PCG screening is 97.99% and 99.42%, respectively. For the five‐category dataset, the average accuracy is 99%.

Deep Learning Algorithms to Detect Murmurs Associated With Structural Heart Disease.

Background The success of cardiac auscultation varies widely among medical professionals, which can lead to missed treatments for structural heart disease. Applying machine learning to cardiac auscultation could address this problem, but despite recent interest, few algorithms have been brought to clinical practice. We evaluated a novel suite of Food and Drug Administration-cleared algorithms trained via deep learning on >15 000 heart sound recordings. Methods and Results We validated the algorithms on a data set of 2375 recordings from 615 unique subjects. This data set was collected in real clinical environments using commercially available digital stethoscopes, annotated by board-certified cardiologists, and paired with echocardiograms as the gold standard. To model the algorithm in clinical practice, we compared its performance against 10 clinicians on a subset of the validation database. Our algorithm reliably detected structural murmurs with a sensitivity of 85.6% and specificity of 84.4%. When limiting the analysis to clearly audible murmurs in adults, performance improved to a sensitivity of 97.9% and specificity of 90.6%. The algorithm also reported timing within the cardiac cycle, differentiating between systolic and diastolic murmurs. Despite optimizing acoustics for the clinicians, the algorithm substantially outperformed the clinicians (average clinician accuracy, 77.9%; algorithm accuracy, 84.7%.) Conclusions The algorithms accurately identified murmurs associated with structural heart disease. Our results illustrate a marked contrast between the consistency of the algorithm and the substantial interobserver variability of clinicians. Our results suggest that adopting machine learning algorithms into clinical practice could improve the detection of structural heart disease to facilitate patient care.

Heart murmur detection from phonocardiogram recordings: The George B. Moody PhysioNet Challenge 2022.

Cardiac auscultation is an accessible diagnostic screening tool that can help to identify patients with heart murmurs, who may need follow-up diagnostic screening and treatment for abnormal cardiac function. However, experts are needed to interpret the heart sounds, limiting the accessibility of cardiac auscultation in resource-constrained environments. Therefore, the George B. Moody PhysioNet Challenge 2022 invited teams to develop algorithmic approaches for detecting heart murmurs and abnormal cardiac function from phonocardiogram (PCG) recordings of heart sounds. For the Challenge, we sourced 5272 PCG recordings from 1452 primarily pediatric patients in rural Brazil, and we invited teams to implement diagnostic screening algorithms for detecting heart murmurs and abnormal cardiac function from the recordings. We required the participants to submit the complete training and inference code for their algorithms, improving the transparency, reproducibility, and utility of their work. We also devised an evaluation metric that considered the costs of screening, diagnosis, misdiagnosis, and treatment, allowing us to investigate the benefits of algorithmic diagnostic screening and facilitate the development of more clinically relevant algorithms. We received 779 algorithms from 87 teams during the Challenge, resulting in 53 working codebases for detecting heart murmurs and abnormal cardiac function from PCG recordings. These algorithms represent a diversity of approaches from both academia and industry, including methods that use more traditional machine learning techniques with engineered clinical and statistical features as well as methods that rely primarily on deep learning models to discover informative features. The use of heart sound recordings for identifying heart murmurs and abnormal cardiac function allowed us to explore the potential of algorithmic approaches for providing more accessible diagnostic screening in resource-constrained environments. The submission of working, open-source algorithms and the use of novel evaluation metrics supported the reproducibility, generalizability, and clinical relevance of the research from the Challenge.

Novel phonocardiography system for heartbeat detection from various locations

The paper presents evaluation of the proposed phonocardiography (PCG) measurement system designed primarily for heartbeat detection to estimate heart rate (HR). Typically, HR estimation is performed using electrocardiography (ECG) or pulse wave as one of the fundamental diagnostic methodologies for assessing cardiac function. The system includes novel both sensory part and data processing procedure, which is based on signal preprocessing using Wavelet Transform (WT) and Shannon energy computation and heart sounds classification using K-means. Due to the lack of standardization in the placement of PCG sensors, the study focuses on evaluating the signal quality obtained from 7 different sensor locations on the subject’s chest and investigates which locations are most suitable for recording heart sounds. The suitability of sensor localization was examined in 27 subjects by detecting the first two heart sounds (S1, S2). The HR detection sensitivity related to reference ECG from all sensor positions reached values over 88.9 and 77.4% in detection of S1 and S2, respectively. The placement in the middle of sternum showed the higher signal quality with median of the proper S1 and S2 detection sensitivity of 98.5 and 97.5%, respectively.

Health Monitoring via Heart, Breath, and Korotkoff Sounds by Wearable Piezoelectret Patches.

Real-time monitoring of vital sounds from cardiovascular and respiratory systems via wearable devices together with modern data analysis schemes have the potential to reveal a variety of health conditions. Here, a flexible piezoelectret sensing system is developed to examine audio physiological signals in an unobtrusive manner, including heart, Korotkoff, and breath sounds. A customized electromagnetic shielding structure is designed for precision and high-fidelity measurements and several unique physiological sound patterns related to clinical applications are collected and analyzed. At the left chest location for the heart sounds, the S1 and S2 segments related to cardiac systole and diastole conditions, respectively, are successfully extracted and analyzed with good consistency from those of a commercial medical device. At the upper arm location, recorded Korotkoff sounds are used to characterize the systolic and diastolic blood pressure without a doctor or prior calibration. An Omron blood pressure monitor is used to validate these results. The breath sound detections from the lung/ trachea region are achieved a signal-to-noise ration comparable to those of a medical recorder, BIOPAC, with pattern classification capabilities for the diagnosis of viable respiratory diseases. Finally, a 6×6 sensor array is used to record heart sounds at different locations of the chest area simultaneously, including the Aortic, Pulmonic, Erb's point, Tricuspid, and Mitral regions in the form of mixed data resulting from the physiological activities of four heart valves. These signals are then separated by the independent component analysis algorithm and individual heart sound components from specific heart valves can reveal their instantaneous behaviors for the accurate diagnosis of heart diseases. The combination of these demonstrations illustrate a new class of wearable healthcare detection system for potentially advanced diagnostic schemes.

Feature selection algorithms highlight the importance of the systolic segment for normal/murmur PCG beat classification

This paper proposes a method using statistical local and global features for classifying healthy and murmur heart sound recordings from phonocardiogram signals. Classification requires features extraction step that converts each signal into a sequence of feature vectors composed of static and dynamic energy coefficients computed from overlapped analysis windows. Firstly, we propose, for each heartbeat, to extract local features from the local consecutive regions (1st Sound, Systole, 2nd Sound, Diastole) and global ones from the global region. For each region, the features are the statistical features (mean and standard deviation) computed on the feature vector sequence plus the duration. Secondly, we propose to select the relevant features using filter approach based on mutual information criteria. The extraction and selection methods are validated using K nearest neighbor and Gaussian Mixture Models as classifiers. The classification system were evaluated on a sub-dataset of the public PASCAL heart sounds classifying challenge. Results showed that 12 features selected using the Max-Relevance Min-Redundancy selection strategy were sufficient to explain the two classes with 94.97% classification rate higher than 92.74% state-of-the-art rate. We also showed this selection strategy helped the system to be robust to the testing phase when using automatic segmentation rather than manual segmentation. This work demonstrates that local systolic segment features are the most relevant for murmur/normal classification, regardless of segmentation methods. It also shows that feature selection algorithms have potential to highlight certain relevant regions in signals, which is useful for aided diagnostic systems and basic research.

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.

Detection of Twin Pregnancies using Fetal Phonocardiogram.

Fetal phonocardiogram (fPCG), or the electronic recording of fetal heart sounds, is a safe and easily available signal that can be used to monitor fetal wellbeing. In the proposed work an attempt is made to identify twin pregnancies using fPCG data recorded from the fetus with 1/3rd power in octave band filtered output as features to train K-Nearest Neighbor (KNN) and support vector machine (SVM) classifiers. The SVM classifier with the quadratic kernel is able to identify singletons and twins with a positive predictive value of 100% and 79.1% respectively. The KNN classifier with k=10 neighbors is able to identify singletons and twins with a positive predictive value of 100% and 81.8% respectively.Clinical Relevance: Identifying twin pregnancies from singleton is an essential clinical protocol followed during late pregnancy as there may be complications like twin-twin transfusion syndrome, selective fetal growth restriction, and preterm labor in twin pregnancy [1], [2]. Ultrasound imaging is the most commonly used technique for twin pregnancy detection, though it is often not affordable or available in rural or low-income populations. Utilization of fPCG in such circumstances has immense clinical potential.