Machine learning based portable device for detection of cardiac abnormality

Abstract

Heart auscultation is a useful method for diagnosis of cardiac diseases using conventional stethoscope. This paper proposes a convenient, efficient and non-invasive device for automatic detection of cardiac abnormalities. The conventional method takes time and involves medical professionals for detection. Diagnosis of cardiac diseases using automatic diagnosis of heart sounds will be very useful in primary health care (PHC) centers where professional help is unavailable. The proposed work presents a machine learning based portable device for real time diagnosis and classification of cardiac diseases using heart sound. The proposed device is a battery operated and standalone device and uses a low cost 1.2 GHz quad core processor and a digital medical instrument to record and listen the heart sound which indicates that the proposed device is efficient and a low cost device making it suitable for applications in real time.