Computerized detection & classification of ECG signals

An electrocardiogram (ECG) is a recorded representation of electrical activity of heart with respect to time. Being a very important diagnostic tool for determining various heart functions, computer based ECG interpretation is a computing system that records and automatically detects the variations in important parameters of a given ECG recording. These parameters are very critical for doctors while examining the condition of the heart of a patient. Identification of the specific area of the heart under problem and the extent of damage done are the basic areas which have been highlighted for proper diagnosis to save patient´s life. This paper, proposes an approach to determine the values of the amplitude of P, QRS & T together with time durations (PR, RR, QRS, and QT). From these calculations, the heart condition whether it is normal or not, is determined along with amount of variation so that the cardiologist can recommend the amount of dosing required for the medicine for the ailment of the patient. It also classifies the ECG into seven major cardiac problems. In this thesis, we have introduced a curve mapping technique based on slope method to detect & classify the ECG signals into seven major cardiac problems. The accuracy rate arrived in the slope method ranges from 92% to 100%, for seven cardiac conditions with an overall average 96.57 %.