Wrist Pulse Classification System for Healthy and Unhealthy Subjects

Abstract

According to Ayurveda, human health status can be measured by using three fingers as 'VATA', 'PITTA' and 'KAPHA' called Tridosha. This is very effective method to know the status of human body. But for this type of diagnosis practitioner requires long term training. Also the diagnoses from practitioner often deviate greatly due to their subjective experience. Thus, it is highly required to design such a system which diagnoses human health status objectively not subjectively. In this paper, design of an embedded system, which classifies the health status, is discussed. This system first convert time domain signal into frequency domain using Fast Fourier Transform (FFT). Using Band Energy Ratio (BER) feature vector is retrieved from frequency domain signal. This feature vector is applied to the linear as well as quadratic classifier. This will classify the human health in a group of two as healthy or unhealthy.