| Summary: | 博士 === 國立成功大學 === 電機工程學系 === 103 === For cardiovascular disease patients, non-invasive diagnosis methods, such as echocardiogram and electrocardiography (ECG), provide an accurate and safe method to assess the heart function. However, the echocardiogram is expensive and requires the operation of trained cardiologist. Although ECG is a standard method to diagnosis the heart disease in the first screening process, it cannot detect the mechanical disorder of heart function and thrombosis of blood vessel. Accordingly, this study proposed a Human-like auditory processing (HAP) to analysis the cardiovascular signal. The HAP had been verified and performed well at speech processing and robust speech recognition. It was further combined with electrical stethoscope and computerized-auscultation method for mimicking a trained practitioner in performing the auscultation process. In the proposed approach, the bruit obtained by a standard phonoangiography and phonocardiography of heart murmur are transformed into the time-frequency domain, and two spectro-temporal features, namely the auditory spectrum flux and the auditory spectral centroid, are then extracted. The distributions of the two features are analyzed using a multivariate Gaussian distribution (MGD) model. Finally, the distribution parameters of the MGD model are used to detect the presence (or otherwise) of vascular access stenosis. The results show an accuracy of 83.87% in detecting significant vascular access stenosis. Besides, the PDA murmurs, are used in the blind test for algorithm effectiveness assessment. The results demonstrate that the proposed computer-assisted auscultation method can achieve a high sensitivity of 100% and a specificity of 91.67% for PDA detection. The above results demonstrate that the proposed human-like auditory processing system used for cardiovascular sound analysis is robust, cost-effective and convenient for the non-invasive early detection.
|
|---|
