Annals of noninvasive electrocardiology : the official journal of the International Society for Holter and Noninvasive Electrocardiology, Inc · 2026
Extraction of Acoustic Features via Empirical Wavelet Transform to Determine Stenosis Degree of the Left Anterior Descending Artery Based on the Diastolic Heart Sounds of 75 Participants.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Li Haixia, Zhang Yafang, Ren Guofeng, Tian Yun, Chai Yan, Wang Xiaoyan
Аннотация
OBJECTIVES: This study aimed to develop a method for extracting acoustic features to assess left anterior descending artery (LAD) stenosis severity.
METHODS: Heart sound data were collected from 75 participants (10 diastoles per participant) using a high-signal-to-noise ratio micro-electro-mechanical systems stethoscope. The diastolic signals were preprocessed, and empirical wavelet transform was applied to decompose their power spectra into three modes (0-150, 150-500, and > 500 Hz). The spectral energies (e(1), e(2), e(3)) of these modes were analyzed, and support vector machine (SVM) and extreme gradient boosting (XGBoost) machine learning algorithms were used to classify LAD stenosis into mild (< 50%), moderate (50%-75%), and severe (> 75%).
RESULTS: Spectral energies e(2) and e(3) significantly increased with stenosis severity, and XGBoost outperformed SVM, achieving a test accuracy of 0.8133 and areas under the curve of 0.9358, 0.9644, and 0.9580 for mild, moderate, and severe stenosis, respectively.
CONCLUSION: Empirical wavelet transform-extracted spectral energies of e(2) and e(3), combined with XGBoost, effectively determine LAD stenosis degree, offering a non-invasive screening tool.
METHODS: Heart sound data were collected from 75 participants (10 diastoles per participant) using a high-signal-to-noise ratio micro-electro-mechanical systems stethoscope. The diastolic signals were preprocessed, and empirical wavelet transform was applied to decompose their power spectra into three modes (0-150, 150-500, and > 500 Hz). The spectral energies (e(1), e(2), e(3)) of these modes were analyzed, and support vector machine (SVM) and extreme gradient boosting (XGBoost) machine learning algorithms were used to classify LAD stenosis into mild (< 50%), moderate (50%-75%), and severe (> 75%).
RESULTS: Spectral energies e(2) and e(3) significantly increased with stenosis severity, and XGBoost outperformed SVM, achieving a test accuracy of 0.8133 and areas under the curve of 0.9358, 0.9644, and 0.9580 for mild, moderate, and severe stenosis, respectively.
CONCLUSION: Empirical wavelet transform-extracted spectral energies of e(2) and e(3), combined with XGBoost, effectively determine LAD stenosis degree, offering a non-invasive screening tool.