An electronic system to recognize heart diseases based on heart sounds: A stochasatic alogorithm implemented on DSK6713

Stethoscope is one of the critical tools used to assess a patient's health by performing auscultation. There is a necessity to develop a device which can recognize the heart disease that a patient might have in the initial stages of examination. This paper proposes a stand-alone system based on DSK6713 that can identify an abnormality in a heart sound while the auscultation is being performed. The technique that has been developed involves sound amplification and analysis techniques. Analysis of the recorded auscultation sounds is performed by using stochastic algorithms, which are tested on a database of recorded heart sounds collected from Michigan Heart Library. The algorithm has been implemented on DSK6713 to make the system stand alone. The analysis includes identification of a probable heart disease. The system's output is recognizing the patient's state of the heart, i.e., diseased or normal - if diseased, what is the possible heart disease and where the abnormality is occurring. The future work includes real time classification of the abnormality using training sets that are recorded on actual patients.