Heart Sound Multiclass Analysis Based on Raw Data and Convolutional Neural Network

Avanzato, Roberta; Beritelli, Francesco

doi:10.1109/lsens.2020.3039366

Cited by 23 publications

(12 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Handcrafted features may cause loss of information for classification, and therefore several studies tried to use 1D convolutions to extract features from a raw data segment. Avanzato and Beritelli 25 built a five-layered 1D CNN 25 to extract features from raw PCG data, while Alkhodari and Fraiwan 27 used a three-layered 1D CNN cascaded by a Bi-directional Long Short-Term Memory (BiLSTM). In this study, we consider directly processing the noise-cancelled PCG segment with a clique block-based DNN, where clique blocks and channel-wise attention are employed.…”

Section: Discussionmentioning

confidence: 99%

“…The first objective of this study is to establish a DNN-based cardiac auscultation method that can automatically assess the severity of MR. Based on the findings of previous studies,23–31 the hypothesis in this study is that putting PCG signals into a trained DNN can lead to automatic classification of the severity of MR, with high sensitivity and specificity. The second objective is to quantitatively measure the performance of the developed artificial intelligence (AI)-based MR assessment method when the AI-aided electronic stethoscope Smartho-D2 is used by cardiologists as well as patients.…”

Section: Introductionmentioning

confidence: 99%

“…The segments were divided evenly into five categories: normal, aortic stenosis, MR, mitral stenosis and mitral valve prolapse. Several works were carried out in handling this five-category classification task 23–29. Time-frequency magnitude and phase features were used,23 and graph-based feature was developed,24 with traditional machine learning methods used for classification.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Developing an AI-assisted digital auscultation tool for automatic assessment of the severity of mitral regurgitation: protocol for a cross-sectional, non-interventional study

Zhang,

Cheng,

et al. 2024

BMJ Open

View full text Add to dashboard Cite

IntroductionMitral regurgitation (MR) is the most common valvular heart disorder, with a morbidity rate of 2.5%. While echocardiography is commonly used in assessing MR, it has many limitations, especially for large-scale MR screening. Cardiac auscultation with electronic stethoscope and artificial intelligence (AI) can be a fast and economical modality for assessing MR severity. Our objectives are (1) to establish a deep neural network (DNN)-based cardiac auscultation method for assessing the severity of MR; and (2) to quantitatively measure the performance of the developed AI-based MR assessment method by virtual clinical trial.Methods and analysisIn a cross-sectional design, phonocardiogram will be recorded at the mitral valve auscultation area of outpatients. The enrolled patients will be checked by echocardiography to confirm the diagnosis of MR or no MR. Echocardiographic parameters will be used as gold standard to assess the severity of MR, classified into four levels: none, mild, moderate and severe. The study consists of two stages. First, an MR-related cardiac sound database will be created on which a DNN-based MR severity classifier will be trained. The automatic MR severity classifier will be integrated with the Smartho-D2 electronic stethoscope. Second, the performance of the developed smart device will be assessed in an independent clinical validation data set. Sensitivity, specificity, precision, accuracy and F1 score of the developed smart MR assessment device will be evaluated. Agreement on the performance of the smart device between cardiologist users and patient users will be inspected. The interpretability of the developed model will also be studied with statistical comparisons of occlusion map-guided variables among the four severity groups.Ethics and disseminationThe study protocol was approved by the Medical Ethics Committee of Huzhou Central Hospital, China (registration number: 202302009-01). Informed consent is required from all participants. Dissemination will be through conference presentations and peer-reviewed journals.Trial registration numberChiCTR2300069496.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Developing an AI-assisted digital auscultation tool for automatic assessment of the severity of mitral regurgitation: protocol for a cross-sectional, non-interventional study

Zhang,

Cheng,

et al. 2024

BMJ Open

View full text Add to dashboard Cite

show abstract

“…In recent years, as time-frequency representations and other features still need human efforts to select features, it has been becoming increasingly popular to use end-to-end networks to learn representations from heart sounds. Based on raw heart sound signals, various 1D CNN architectures have been proposed and applied to the task of heart sound classification [10], [34], [39], [51], [127]. Furthermore, Liu et al introduced a temporal convolutional network (TCN) that performed a high sensitivity for heart sound classification [74], as a TCN benefiting from dilated and casual convolutions is more suitable for sequential data than typical CNNs are.…”

Section: B Deep Learning For Classificationmentioning

confidence: 99%

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

Zhao¹,

Chen²,

Nguyên³

et al. 2023

Preprint

View full text Add to dashboard Cite

Heart sound auscultation has been demonstrated to be beneficial in clinical usage for early screening of cardiovascular diseases. Due to the high requirement of well-trained professionals for auscultation, automatic auscultation benefiting from signal processing and machine learning can help auxiliary diagnosis and reduce the burdens of training professional clinicians. Nevertheless, classic machine learning is limited to performance improvement in the era of big data. Deep learning has achieved better performance than classic machine learning in many research fields, as it employs more complex model architectures with stronger capability of extracting effective representations. Deep learning has been successfully applied to heart sound analysis in the past years. As most review works about heart sound analysis were given before 2017, the present survey is the first to work on a comprehensive overview to summarise papers on heart sound analysis with deep learning in the past six years 2017-2022. We introduce both classic machine learning and deep learning for comparison, and further offer insights about the advances and future research directions in deep learning for heart sound analysis.

show abstract

“…Meanwhile, two metrics are evaluated, that is, psychology recognition accuracy and psychology recognition time. At the first part, other three neural network structures (ie, RNN, 7 CNN 8 and GNN 9 ) are used as the baselines. At the second part, the recognition method without considering ICN paradigm is used as the baseline.…”

Section: Simulationsmentioning

confidence: 99%

Information‐centric networking driven psychology recognition and health conduction of undergraduates

Zhang

2021

Internet Technology Letters

View full text Add to dashboard Cite

It is considerably significant to recognize the psychology of undergraduates especially during the period of COVID‐19 outbreak. This letter proposes an Information‐Centric Networking (ICN) driven psychology recognition and health conduction method based on the active data collection mode, that is, all undergraduates submit their psychology data to manger without protest by the special sensor tool. Meanwhile, the involved undergraduates and the management center construct a large network model. In order to support the mobility of undergraduates, this letter exploits ICN paradigm to help the management and transmission of psychology data. In addition, this management center deploys Deep Belief Network (DBN) to analyze the submitted psychology data. Finally, this letter makes a practical simulation on campus, and the experimental results show that the proposed psychology control method not only has good recognition effect but also contributes to health conduction. To be specific, the proposed method can improve about 2.6% psychology recognition accuracy and decrease about 46.5% psychology recognition time.

show abstract

Heart Sound Multiclass Analysis Based on Raw Data and Convolutional Neural Network

Cited by 23 publications

References 14 publications

Developing an AI-assisted digital auscultation tool for automatic assessment of the severity of mitral regurgitation: protocol for a cross-sectional, non-interventional study

Developing an AI-assisted digital auscultation tool for automatic assessment of the severity of mitral regurgitation: protocol for a cross-sectional, non-interventional study

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

Information‐centric networking driven psychology recognition and health conduction of undergraduates

Contact Info

Product

Resources

About