Heart Sound Classification Network Based on Convolution and Transformer

Multi-level feature encoding algorithm based on FBPSI for heart sound classification

Analysis of heart sound signals plays an essential role in preventing and diagnosing cardiac diseases. This study proposes a multi-level feature encoding algorithm based on frequency-balanced power spectral intensity for heart sound signal classification. Firstly, a wavelet threshold function is employed to denoise the heart sound signals. Then, the frequency-balanced power spectral intensity envelope is calculated, and an encoder is utilized to extract multi-level features based on the envelope. Finally, an ensemble bagging tree classifier is selected for classification. The experimental data includes binary classification data from the 2016 PhysioNet/CinC Challenge and ternary classification data from the self-collected hypertrophic cardiomyopathy dataset. Results demonstrate that the proposed algorithm performs well, achieving an average classification accuracy of 98.73% for normal and abnormal heart sounds, and 98.12% for normal and two types of hypertrophic cardiomyopathy heart sounds. The proposed method holds significant reference value for the early diagnosis of heart diseases.

Prediction of Arteriovenous Access Dysfunction by Mel Spectrogram-based Deep Learning Model

Background: The early detection of arteriovenous (AV) access dysfunction is crucial for maintaining the patency of vascular access. This study aimed to use deep learning to predict AV access malfunction necessitating further vascular management. Methods: This prospective cohort study enrolled prevalent hemodialysis (HD) patients with an AV fistula or AV graft from a single HD center. Their AV access bruit sounds were recorded weekly using an electronic stethoscope from three different sites (arterial needle site, venous needle site, and the midpoint between the arterial and venous needle sites) before HD sessions. The audio signals were converted to Mel spectrograms using Fourier transformation and utilized to develop deep learning models. Three deep learning models, (1) Convolutional Neural Network (CNN), (2) Convolutional Recurrent Neural Network (CRNN), and (3) Vision Transformers-Gate Recurrent Unit (ViT-GRU), were trained and compared to predict the likelihood of dysfunctional AV access. Results: Total 437 audio recordings were obtained from 84 patients. The CNN model outperformed the other models in the test set, with an F1 score of 0.7037 and area under the receiver operating characteristic curve (AUROC) of 0.7112. The Vit-GRU model had high performance in out-of-fold predictions, with an F1 score of 0.7131 and AUROC of 0.7745, but low generalization ability in the test set, with an F1 score of 0.5225 and AUROC of 0.5977. Conclusions: The CNN model based on Mel spectrograms could predict malfunctioning AV access requiring vascular intervention within 10 days. This approach could serve as a useful screening tool for high-risk AV access.

Detection of Coronary Artery Disease Based on Clinical Phonocardiogram and Multiscale Attention Convolutional Compression Network

Heart sound is an important physiological signal that contains rich pathological information related with coronary stenosis. Thus, some machine learning methods are developed to detect coronary artery disease (CAD) based on phonocardiogram (PCG). However, current methods lack sufficient clinical dataset and fail to achieve efficient feature utilization. Besides, the methods require complex processing steps including empirical feature extraction and classifier design. To achieve efficient CAD detection, we propose the multiscale attention convolutional compression network (MACCN) based on clinical PCG dataset. Firstly, PCG dataset including 102 CAD subjects and 82 non-CAD subjects was established. Then, a multiscale convolution structure was developed to catch comprehensive heart sound features and a channel attention module was developed to enhance key features in multiscale attention convolutional block (MACB). Finally, a separate downsampling block was proposed to reduce feature losses. MACCN combining the blocks can automatically extract features without empirical and manual feature selection. It obtains good classification results with accuracy 93.43%, sensitivity 93.44%, precision 93.48%, and F1 score 93.42%. The study implies that MACCN performs effective PCG feature mining aiming for CAD detection. Further, it integrates feature extraction and classification and provides a simplified PCG processing case.

A Noise-Robust Heart Sound Segmentation Algorithm Based on Shannon Energy

Heart sound segmentation has been shown to improve the performance of artificial intelligence (AI)-based auscultation decision support systems increasingly viewed as a solution to compensate for eroding auscultatory skills and the associated subjectivity. Various segmentation approaches with demonstrated performance can be utilized for this task, but their robustness can suffer in the presence of noise. A noise-robust heart sound segmentation algorithm was developed and its accuracy was tested using two datasets: the CirCor DigiScope Phonocardiogram dataset and an in-house dataset - a heart murmur library collected at the Children's National Hospital (CNH). On the CirCor dataset, our segmentation algorithm marked the boundaries of the primary heart sounds S1 and S2 with an accuracy of 0.28 ms and 0.29 ms, respectively, and correctly identified the actual positive segments with a sensitivity of 97.44%. The algorithm also executed four times faster than a logistic regression hidden semi-Markov model. On the CNH dataset, the algorithm succeeded in 87.4% cases, achieving a 6% increase in segmentation success rate demonstrated by our original Shannon energy-based algorithm. Accurate heart sound segmentation is critical to supporting and accelerating AI research in cardiovascular diseases. The proposed algorithm increases the robustness of heart sound segmentation to noise and viability for clinical use.