Quantitative analysis of heart sounds and systolic heart murmurs using wavelet transform and AR modeling

Heart Murmur Classification for Diagnostic Applications

This paper presents an innovative methodology for applying signal processing techniques and machine learning to process recently available data for heart murmur detection and classification. The scope of this research is to improve upon current heart murmur detection research by investigating the classification of murmur characteristics, such as pitch, duration, configuration, and quality, which are necessary for diagnosis. Feature extraction was performed on each cardiac cycle for a total of 53 features. The machine learning models tested included 5 supervised models, with 3 of those being boosting models, and 13 voting classifier combinations. The efficiency of the feature extraction methodology and the machine learning algorithms was evaluated with a confusion matrix. The murmur detection algorithm achieved an F-score of 0.87.

Heart sound segmentation of pediatric auscultations using wavelet analysis

Auscultation is widely applied in clinical activity, nonetheless sound interpretation is dependent on clinician training and experience. Heart sound features such as spatial loudness, relative amplitude, murmurs, and localization of each component may be indicative of pathology. In this study we propose a segmentation algorithm to extract heart sound components (S1 and S2) based on it's time and frequency characteristics. This algorithm takes advantage of the knowledge of the heart cycle times (systolic and diastolic periods) and of the spectral characteristics of each component, through wavelet analysis. Data collected in a clinical environment, and annotated by a clinician was used to assess algorithm's performance. Heart sound components were correctly identified in 99.5% of the annotated events. S1 and S2 detection rates were 90.9% and 93.3% respectively. The median difference between annotated and detected events was of 33.9 ms.