An Efficient Signal Processing Technique for Automated Myocardial Infarction Detection

Electrocardiography(ECG) is a non-invasive diagnostic tool for diagnosing Cardiovascular disease such as coronary artery disease. ECG is a recording of electrical activity of heart lying in mV range so the process of manual detection of heart disease is challenging, consumes time and liable to human errors. Thus an automatic detection of Myocardial Infarction(MI) is done by using single lead(V3) with variable length beat for classifying all eleven classes of MI. In this paper, the Discrete Wavelet Transform(DWT) coefficient's are computed for each heart beat and after that Principal Component Analysis(PCA) technique is deployed for minimizing the number of coefficients. DWT coefficient's along with PCA is deployed for optimum feature extraction, further Support Vector Machine(SVM) with Radial Basis Function(RBF) kernel is utilized for classification of 11 types of MI and healthy control. The experiment are carried out in MATLAB using PTB ECG diagnostic database. The proposed method gave best accuracy as compared with existing work of 99.02% for classifying 12 classes.