Machine Learning Algorithms for Processing and Classifying Unsegmented Phonocardiographic Signals: An Efficient Edge Computing Solution Suitable for Wearable Devices.

The phonocardiogram (PCG) can be used as an affordable way to monitor heart conditions. This study proposes the training and testing of several classifiers based on SVMs (support vector machines), k-NN (k-Nearest Neighbor), and NNs (neural networks) to perform binary ("Normal"/"Pathologic") and multiclass ("Normal", "CAD" (coronary artery disease), "MVP" (mitral valve prolapse), and "Benign" (benign murmurs)) classification of PCG signals, without heart sound segmentation algorithms. Two datasets of 482 and 826 PCG signals from the Physionet/CinC 2016 dataset are used to train the binary and multiclass classifiers, respectively. Each PCG signal is pre-processed, with spike removal, denoising, filtering, and normalization; afterward, it is divided into 5 s frames with a 1 s shift. Subsequently, a feature set is extracted from each frame to train and test the binary and multiclass classifiers. Concerning the binary classification, the trained classifiers yielded accuracies ranging from 92.4 to 98.7% on the test set, with memory occupations from 92.7 kB to 11.1 MB. Regarding the multiclass classification, the trained classifiers achieved accuracies spanning from 95.3 to 98.6% on the test set, occupying a memory portion from 233 kB to 14.1 MB. The NNs trained and tested in this work offer the best trade-off between performance and memory occupation, whereas the trained k-NN models obtained the best performance at the cost of large memory occupation (up to 14.1 MB). The classifiers' performance slightly depends on the signal quality, since a denoising step is performed during pre-processing. To this end, the signal-to-noise ratio (SNR) was acquired before and after the denoising, indicating an improvement between 15 and 30 dB. The trained and tested models occupy relatively little memory, enabling their implementation in resource-limited systems.

The Evolving Stethoscope: Insights Derived from Studying Phonocardiography in Trainees.

Phonocardiography (PCG) is used as an adjunct to teach cardiac auscultation and is now a function of PCG-capable stethoscopes (PCS). To evaluate the efficacy of PCG and PCS, the authors investigated the impact of providing PCG data and PCSs on how frequently murmurs, rubs, and gallops (MRGs) were correctly identified by third-year medical students. Following their internal medicine rotation, third-year medical students from the Georgetown University School of Medicine completed a standardized auscultation assessment. Sound files of 10 different MRGs with a corresponding clinical vignette and physical exam location were provided with and without PCG (with interchangeable question stems) as 10 paired questions (20 total questions). Some (32) students also received a PCS to use during their rotation. Discrimination/difficulty indexes, comparative chi-squared, and McNemar test p-values were calculated. The addition of phonocardiograms to audio data was associated with more frequent identification of mitral stenosis, S4, and cardiac friction rub, but less frequent identification of ventricular septal defect, S3, and tricuspid regurgitation. Students with a PCS had a higher frequency of identifying a cardiac friction rub. PCG may improve the identification of low-frequency, usually diastolic, heart sounds but appears to worsen or have little effect on the identification of higher-frequency, often systolic, heart sounds. As digital and phonocardiography-capable stethoscopes become more prevalent, insights regarding their strengths and weaknesses may be incorporated into medical school curricula, bedside rounds (to enhance teaching and diagnosis), and telemedicine/tele-auscultation efforts.

Emerging Paradigms in Fetal Heart Rate Monitoring: Evaluating the Efficacy and Application of Innovative Textile-Based Wearables.

This comprehensive review offers a thorough examination of fetal heart rate (fHR) monitoring methods, which are an essential component of prenatal care for assessing fetal health and identifying possible problems early on. It examines the clinical uses, accuracy, and limitations of both modern and traditional monitoring techniques, such as electrocardiography (ECG), ballistocardiography (BCG), phonocardiography (PCG), and cardiotocography (CTG), in a variety of obstetric scenarios. A particular focus is on the most recent developments in textile-based wearables for fHR monitoring. These innovative devices mark a substantial advancement in the field and are noteworthy for their continuous data collection capability and ergonomic design. The review delves into the obstacles that arise when incorporating these wearables into clinical practice. These challenges include problems with signal quality, user compliance, and data interpretation. Additionally, it looks at how these technologies could improve fetal health surveillance by providing expectant mothers with more individualized and non-intrusive options, which could change the prenatal monitoring landscape.

Detection of valvular heart diseases combining orthogonal non-negative matrix factorization and convolutional neural networks in PCG signals.

BACKGROUND AND OBJECTIVE: Valvular heart disease (VHD) is associated with elevated mortality rates. Although transthoracic echocardiography (TTE) is the gold standard detection tool, phonocardiography (PCG) could be an alternative as it is a cost-effective and noninvasive method for cardiac auscultation. Many researchers have dedicated their efforts to improving the decision-making process and developing robust and precise approaches to assist physicians in providing reliable diagnoses of VHD. METHODS: This research proposes a novel approach for the detection of anomalous valvular heart sounds from PCG signals. The proposed approach combines orthogonal non-negative matrix factorization (ONMF) and convolutional neural network (CNN) architectures in a three-stage cascade. The aim of the proposal is to improve the learning process by identifying the optimal ONMF temporal or spectral patterns for accurate detection. In the first stage, the time-frequency representation of the input PCG signal is computed. Next, band-pass filtering is performed to locate the spectral range that is most relevant for the presence of such cardiac abnormalities. In the second stage, the temporal and spectral cardiac structures are extracted using the ONMF approach. These structures are utilized in the third stage and fed into the CNN architecture to detect abnormal heart sounds. RESULTS: Several state-of-the-art CNN architectures, such as LeNet5, AlexNet, ResNet50, VGG16 and GoogLeNet, have been evaluated to determine the effectiveness of using ONMF temporal features for VHD detection. The results reveal that the integration of ONMF temporal features with a CNN classifier significantly improve VHD detection. Specifically, the proposed approach achieves an accuracy improvement of approximately 45% when ONMF spectral features are used and 35% when time-frequency features from the short-time Fourier transform (STFT) spectrogram are used. Additionally, feeding ONMF temporal features into low-complexity CNN architectures yields competitive results comparable to those obtained with complex architectures. CONCLUSIONS: The temporal structure factorized by ONMF plays a critical role in distinguishing between normal heart sounds and abnormal heart sounds since the repeatability of normal heart cycles is disrupted by the presence of cardiac abnormalities. Consequently, the results highlight the importance of appropriate input data representation in the learning process of CNN models in the biomedical field of valvular heart sound detection.

Simple Phonocardiography with the Ankle-Brachial Index Measurement System as a Tool for Detecting Fourth Heart Sounds of Hypertrophic Cardiomyopathy.

Conventional phonocardiography is useful for objective assessment of cardiac auscultation, but its availability is limited. More recently, an ankle-brachial index (ABI) measurement system equipped with simple phonocardiography has become widely used for diagnosing peripheral artery disease, however, whether this simple phonocardiography can be an alternative to conventional phonocardiography remains unclear.This retrospective study consisted of 48 patients with hypertrophic cardiomyopathy (HCM) and 107 controls. The presence of the fourth sound (S4) was assessed by conventional phonocardiography, in addition to apexcardiography and auscultation, in all patients with HCM. S4 was also estimated by the ABI measurement system with the phonocardiographic microphone on the sternum (the standard method) or at the apex (the apex method) in HCM patients and controls.S4 on conventional phonocardiography was detected in 42 of 48 patients (88%) with HCM. Auscultation for the detection of S4 had a sensitivity of 0.78, specificity of 0.57, and accuracy of 0.75. These diagnostic values were generally superior to those of the standard method using the ABI measurement system, whereas the apex method using the ABI measurement system had better diagnostic values, with an excellent specificity of 1.0, sensitivity of 0.77, and accuracy of 0.80. No significant differences were observed in low ABI defined as < 0.9.Simple phonocardiography equipped with the ABI measurement system may be an alternative to conventional phonocardiography for the detection of S4 in patients with HCM when the phonocardiographic microphone is moved from the sternum to the apex.

A Robust and Real-Time Capable Envelope-Based Algorithm for Heart Sound Classification: Validation under Different Physiological Conditions.

This paper proposes a robust and real-time capable algorithm for classification of the firstand second heart sounds. The classification algorithm is based on the evaluation of the envelope curveof the phonocardiogram. For the evaluation, in contrast to other studies, measurements on twelveprobands were conducted in different physiological conditions. Moreover, for each measurement theauscultation point, posture and physical stress were varied. The proposed envelope-based algorithmis tested with two different methods for envelope curve extraction: the Hilbert transform andthe short-time Fourier transform. The performance of the classification of the first heart soundsis evaluated by using a reference electrocardiogram. Overall, by using the Hilbert transform,the algorithm has a better performance regarding the F1-score and computational effort. Theproposed algorithm achieves for the S1 classification an F1-score up to 95.7% and in average 90.5 %.The algorithm is robust against the age, BMI, posture, heart rate and auscultation point (exceptmeasurements on the back) of the subjects. The ECG and PCG records are available from the authors.

A Wavelet Transform-Based Neural Network Denoising Algorithm for Mobile Phonocardiography.

Cardiovascular pathologies cause 23.5% of human deaths, worldwide. An auto-diagnostic system monitoring heart activity, which can identify the early symptoms of cardiac illnesses, might reduce the death rate caused by these problems. Phonocardiography (PCG) is one of the possible techniques able to detect heart problems. Nevertheless, acoustic signal enhancement is required since it is exposed to various disturbances coming from different sources. The most common denoising enhancement is based on the Wavelet Transform (WT). However, the WT is highly susceptible to variations in the noise frequency distribution. This paper proposes a new adaptive denoising algorithm, which combines WT and Time Delay Neural Networks (TDNN). The acquired signal is decomposed by means of the WT using the coif five-wavelet basis at the tenth decomposition level and then provided as input to the TDNN. Besides the advantage of adaptive thresholding, the reason for using TDNNs is their capacity of estimating the Inverse Wavelet Transform (IWT). The best parameters of the TDNN were found for a NN consisting of 25 neurons in the first and 15 in the second layer and the delay block of 12 samples. The method was evaluated on several pathological heart sounds and on signals recorded in a noisy environment. The performance of the developed system with respect to other wavelet-based denoising approaches was validated by the online questionnaire.

A Novel Method for Measuring the Timing of Heart Sound Components through Digital Phonocardiography.

The auscultation of heart sounds has been for decades a fundamental diagnostic tool in clinical practice. Higher effectiveness can be achieved by recording the corresponding biomedical signal, namely the phonocardiographic signal, and processing it by means of traditional signal processing techniques. An unavoidable processing step is the heart sound segmentation, which is still a challenging task from a technical viewpoint-a limitation of state-of-the-art approaches is the unavailability of trustworthy techniques for the detection of heart sound components. The aim of this work is to design a reliable algorithm for the identification and the classification of heart sounds' main components. The proposed methodology was tested on a sample population of 24 healthy subjects over 10-min-long simultaneous electrocardiographic and phonocardiographic recordings and it was found capable of correctly detecting and classifying an average of 99.2% of the heart sounds along with their components. Moreover, the delay of each component with respect to the corresponding R-wave peak and the delay among the components of the same heart sound were computed: the resulting experimental values are coherent with what is expected from the literature and what was obtained by other studies.

Classifying Heart Sounds Using Images of Motifs, MFCC and Temporal Features.

Cardiovascular disease is the leading cause of death in the world, and its early detection is a key to improving long-term health outcomes. The auscultation of the heart is still an important method in the medical process because it is very simple and cheap. To detect possible heart anomalies at an early stage, an automatic method enabling cardiac health low-cost screening for the general population would be highly valuable. By analyzing the phonocardiogram signals, it is possible to perform cardiac diagnosis and find possible anomalies at an early-term. Therefore, the development of intelligent and automated analysis tools of the phonocardiogram is very relevant. In this work, we use simultaneously collected electrocardiograms and phonocardiograms from the Physionet Challenge database with the main objective of determining whether a phonocardiogram corresponds to a "normal" or "abnormal" physiological state. Our main contribution is the methodological combination of time domain features and frequency domain features of phonocardiogram signals to improve cardiac disease automatic classification. This novel approach is developed using both features. First, the phonocardiogram signals are segmented with an algorithm based on a logistic regression hidden semi-Markov model, which uses electrocardiogram signals as a reference. Then, two groups of features from the time and frequency domain are extracted from the phonocardiogram segments. One group is based on motifs and the other on Mel-frequency cepstral coefficients. After that, we combine these features into a two-dimensional time-frequency heat map representation. Lastly, a binary classifier is applied to both groups of features to learn a model that discriminates between normal and abnormal phonocardiogram signals. In the experiments, three classification algorithms are used: Support Vector Machines, Convolutional Neural Network, and Random Forest. The best results are achieved when both time and Mel-frequency cepstral coefficients features are considered using a Support Vector Machines with a radial kernel.

Novel algorithm to screen for heart murmurs using computer-aided auscultation in neonates: a prospective single center pilot observational study.

BACKGROUND: Automated detection of heart murmurs with computer-aided auscultation is not yet in clinical routine use. Aim of this study was to test sensitivity and specificity of a novel prototype algorithm in automated detection of heart murmurs from digitally recorded phonocardiograms in neonates admitted at the Neonatal Intensive Care Unit. METHODS: In a prospective pilot observational study from November 2012 to December 2013 auscultations by pediatricians and computer aided auscultation were performed within 12 hours of neonatal echocardiography. Echocardiography was defined as pathological when resulting in any clinical consequences or causing murmur. Phonocardiograms and auscultation were defined as pathological if a murmur was detected. Phonocardiograms were analyzed offline with a novel algorithm prototype (CSD Labs, Graz, Austria) for detection of murmurs in neonates in a first run and with an optimized algorithm in a second run and were compared with echocardiography. Sensitivity and specificity of auscultation by pediatrician and computer aided auscultation were analyzed. RESULTS: Thirty-six neonates (gestational age: 36±3 weeks) were included. Twenty-three (64%) neonates had pathological or murmur causing findings in echocardiography (positive echocardiography). Sensitivity and specificity of auscultation by pediatrician were 17% and 100%, respectively. In comparison to auscultation by pediatrician sensitivity of first run and second run were significantly higher with 70% and 83%, respectively. Specificity of first run and second run were 77% and 85%, respectively. CONCLUSIONS: Phonocardiogram analysis using the novel algorithm prototype had a higher sensitivity than auscultation by pediatrician in detecting positive echocardiography findings in neonates.

Murmur identification and outcome prediction in phonocardiograms using deep features based on Stockwell transform.

Traditionally, heart murmurs are diagnosed through cardiac auscultation, which requires specialized training and experience. The purpose of this study is to predict patients' clinical outcomes (normal or abnormal) and identify the presence or absence of heart murmurs using phonocardiograms (PCGs) obtained at different auscultation points. A semi-supervised model tailored to PCG classification is introduced in this study, with the goal of improving performance using time-frequency deep features. The study begins by investigating the behavior of PCGs in the time-frequency domain, utilizing the Stockwell transform to convert the PCG signal into two-dimensional time-frequency maps (TFMs). A deep network named AlexNet is then used to derive deep feature sets from these TFMs. In feature reduction, redundancy is eliminated and the number of deep features is reduced to streamline the feature set. The effectiveness of the extracted features is evaluated using three different classifiers using the CinC/Physionet challenge 2022 dataset. For Task I, which focuses on heart murmur detection, the proposed approach achieved an average accuracy of 93%, sensitivity of 91%, and F1-score of 91%. According to Task II of the CinC/Physionet challenge 2022, the approach showed a clinical outcome cost of 5290, exceeding the benchmark set by leading methods in the challenge.

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.

Newly identified Phonocardiography frequency bands for psychological stress detection with Deep Wavelet Scattering Network.

The timely psychological stress detection can improve the quality of human life by preventing stress-induced behavioral and pathological consequences. This paper presents a novel framework that eliminates the need of Electrocardiography (ECG) signals-based referencing of Phonocardiography (PCG) signals for psychological stress detection. This stand-alone PCG-based methodology uses wavelet scattering approach on the data acquired from twenty-eight healthy adult male and female subjects to detect psychological stress. The acquired PCG signals are asynchronously segmented for the analysis using wavelet scattering transform. After the noise bands removal, the optimized segmentation length (L), scattering network parameters namely-invariance scale (J) and quality factor (Q) are utilized for computation of scattering features. These scattering coefficients generated are fed to K-nearest neighbor (KNN) and Extreme Gradient Boosting (XGBoost) classifier and the ten-fold cross validation-based performance metrics obtained are-accuracy 94.30 %, sensitivity 97.96 %, specificity 88.01 % and area under the curve (AUC) 0.9298 using XGBoost classifier for detecting psychological stress. Most importantly, the framework also identified two frequency bands in PCG signals with high discriminatory power for psychological stress detection as 270-290 Hz and 380-390 Hz. The elimination of multi-modal data acquisition and analysis makes this approach cost-efficient and reduces computational complexity.

Effects of precise cardio sounds on the success rate of phonocardiography.

This work investigates whether inclusion of the low-frequency components of heart sounds can increase the accuracy, sensitivity and specificity of diagnosis of cardiovascular disorders. We standardized the measurement method to minimize changes in signal characteristics. We used the Continuous Wavelet Transform to analyze changing frequency characteristics over time and to allocate frequencies appropriately between the low-frequency and audible frequency bands. We used a Convolutional Neural Network (CNN) and deep-learning (DL) for image classification, and a CNN equipped with long short-term memory to enable sequential feature extraction. The accuracy of the learning model was validated using the PhysioNet 2016 CinC dataset, then we used our collected dataset to show that incorporating low-frequency components in the dataset increased the DL model's accuracy by 2% and sensitivity by 4%. Furthermore, the LSTM layer was 0.8% more accurate than the dense layer.

Ultrasound phonocardiography for detecting thrombotic formations on bileaflet mechanical heart valves.

BACKGROUND AND AIM OF THE STUDY: The main disadvantage of a mechanical heart valve (MHV) is thrombosis, a serious complication that is associated with high morbidity and mortality. The early detection of thrombotic formations is crucial for a prompt diagnosis and correct therapy before critical symptoms appear in patients. The present study describes the in-vitro assessment of thrombotic deposits by ultrasound phonocardiography on five commercially available bileaflet MHVs. METHODS: The closing sounds produced by bileaflet MHVs were acquired in the frequency range from 6 to 55 kHz. The corresponding power spectra were calculated and then analyzed by an artificial neural network (ANN) trained to classify the presence of simulated thrombotic formations of different weight and shape. Simulations were performed in a Sheffield pulse duplicator under different hydrodynamic regimes. RESULTS: Classification performances of the ANN depend on the range of frequency considered: better performances (up to 100% correct classification) are achieved when the entire spectrum is considered, rather than the audible (down to 87%) and ultrasound (down to 61%) regions, separately. CONCLUSION: Good and very good classification performances are achieved in vitro when phonocardiography is applied to detect and analyze the closing sounds produced by MHVs. Interestingly, extension of the analysis to the ultrasound region can improve classification efficiency. This finding allows the consideration of potential clinical applications of the proposed method to assign an MHV recipient to a risk class, thus enabling a prompt diagnosis.

Design methodology of a new wavelet basis function for fetal phonocardiographic signals.

Fetal phonocardiography (fPCG) based antenatal care system is economical and has a potential to use for long-term monitoring due to noninvasive nature of the system. The main limitation of this technique is that noise gets superimposed on the useful signal during its acquisition and transmission. Conventional filtering may result into loss of valuable diagnostic information from these signals. This calls for a robust, versatile, and adaptable denoising method applicable in different operative circumstances. In this work, a novel algorithm based on wavelet transform has been developed for denoising of fPCG signals. Successful implementation of wavelet theory in denoising is heavily dependent on selection of suitable wavelet basis function. This work introduces a new mother wavelet basis function for denoising of fPCG signals. The performance of newly developed wavelet is found to be better when compared with the existing wavelets. For this purpose, a two-channel filter bank, based on characteristics of fPCG signal, is designed. The resultant denoised fPCG signals retain the important diagnostic information contained in the original fPCG signal.

A proposed novel method for CHD screening by fetal heart murmur detection using phonocardiography.

The paper presents a novel screening method to indicate congenital heart diseases (CHD) which otherwise would remain undetected because of their low level. Therefore, not belonging to the high-risk population, they are omitted from the regular fetal monitoring with ultrasound echocardiography. Based on the fact that CHDs are morphological defects of the heart causing turbulent blood flow, this turbulence appears as a murmur, which can be detected by phonocardiography (PCG). The proposed method applies measurements on the maternal abdomen and from the recorded sound signal a sophisticated processing determines the fetal heart murmur. The paper describes the problems and the additional advantages of the PCG method including the possibility of measurements at home and its combination with the prescribed regular cardiotocographic (CTG) monitoring. The proposed screening process implemented on a telemedicine system provides an enhanced safety against hidden cardiac diseases.

[Research on biometric method of heart sound signal based on GMM].

OBJECTIVE: Extraction of cepstral coefficients combined with Gaussian Mixture Model (GMM) is used to propose a biometric method based on heart sound signal. METHODS: Firstly, the original heart sounds signal was preprocessed by wavelet denoising. Then, Linear Prediction Cepstral Coefficients (LPCC) and Mel Frequency Cepstral Coefficients (MFCC) are compared to extract representative features and develops hidden Markov model (HMM) for signal classification. At last, the experiment collects 100 heart sounds from 50 people to test the proposed algorithm. RESULTS: The comparative experiments prove that LPCC is more suitable than MFCC for heart sound biometric, and by wavelet denoising in each piece of heart sound signal, the system achieves higher recognition rate than traditional GMM. CONCLUSION: Those results show that this method can effectively improve the recognition performance of the system and achieve a satisfactory effect.

A Review of Recent Advances and Future Developments in Fetal Phonocardiography.

Fetal phonocardiography (fPCG) is receiving attention as it is a promising method for continuous fetal monitoring due to its non-invasive and passive nature. However, it suffers from the interference from various sources, overlapping the desired signal in the time and frequency domains. This paper introduces the state-of-the-art methods used for fPCG signal extraction and processing, as well as means of detection and classification of various features defining fetal health state. It also provides an extensive summary of remaining challenges, along with the practical insights and suggestions for the future research directions.

Beyond Heart Murmur Detection: Automatic Murmur Grading From Phonocardiogram.

OBJECTIVE: Murmurs are abnormal heart sounds, identified by experts through cardiac auscultation. The murmur grade, a quantitative measure of the murmur intensity, is strongly correlated with the patient's clinical condition. This work aims to estimate each patient's murmur grade (i.e., absent, soft, loud) from multiple auscultation location phonocardiograms (PCGs) of a large population of pediatric patients from a low-resource rural area. METHODS: The Mel spectrogram representation of each PCG recording is given to an ensemble of 15 convolutional residual neural networks with channel-wise attention mechanisms to classify each PCG recording. The final murmur grade for each patient is derived based on the proposed decision rule and considering all estimated labels for available recordings. The proposed method is cross-validated on a dataset consisting of 3456 PCG recordings from 1007 patients using a stratified ten-fold cross-validation. Additionally, the method was tested on a hidden test set comprised of 1538 PCG recordings from 442 patients. RESULTS: The overall cross-validation performances for patient-level murmur gradings are 86.3% and 81.6% in terms of the unweighted average of sensitivities and F1-scores, respectively. The sensitivities (and F1-scores) for absent, soft, and loud murmurs are 90.7% (93.6%), 75.8% (66.8%), and 92.3% (84.2%), respectively. On the test set, the algorithm achieves an unweighted average of sensitivities of 80.4% and an F1-score of 75.8%. CONCLUSIONS: This study provides a potential approach for algorithmic pre-screening in low-resource settings with relatively high expert screening costs. SIGNIFICANCE: The proposed method represents a significant step beyond detection of murmurs, providing characterization of intensity, which may provide an enhanced classification of clinical outcomes.