Diagnosis of Cardiac Disorders Featuring Pulse Plethysmograph Signals

摘要

Pulse Plethysmography (PuPG) is a technique used to obtain blood volumetric changes. This paper focuses on the acquisition of these signals using PTN-104. The data set was acquired for various Cardiovascular Diseases comprising of Acute Coronary Syndrome, Dilated Cardiomyopathy Calcific Aortic Valve Disease, and Normal people. The data was acquired from the index finger and a total of 1130 signals were gathered comprising of these four classes. The signals were subjected to pre-processing using an amalgam technique comprising of Empirical mode decomposition (EMD) and Discrete Wavelet Transform (DWT). Features were extracted from the pre-processed signal giving a feature vector comprising of the most optimized features. The feature vector was implicated on Support Vector Machines (SVM) for the classification of these four classes giving an accuracy of 99.60% with specificity and sensitivity of 100% and 98.75% respectively.