Hard and soft computing methods for capturing and processing phonocardiogram

Cardiovascular diseases are the biggest cause of deaths worldwide. Heart auscultation based on stethoscope is a noninvasive and a very low cost investigation approach that physician uses to evaluate diseases. To improve auscultation and diagnosis capabilities, a digital stethoscope and a set of digital audio signal processing algorithms have been developed to process adaptively sounds acquired from a couple of microphones embedded in the stethoscope head. Hard computing methods have been applied to the captured digital audio signal to detect cardiovascular diseases. Soft computing inference, such as fuzzy logic, is then proposed to reduce the computational burden of the automatic diseases identification process and to extend the method to the automatic detection of the physiological status of the subject. Finally multimodality and data fusion have been evaluated as methods to improve the diagnostic and identification of the system’s capability.