Classification of Myocardial Infarction Using Multi Resolution Wavelet Analysis of ECG

AbstractIn this paper, classification of anterior and inferior myocardial infarction from normal cases is done using the changes happening in ECG waves. Depth of Q peak and elevation in ST segment is taken in consideration for classification purpose. A multiresolution approach along with an adaptive thresholding is used to extract these ECG features. Classification of inferior myocardial infarction (IMI) and anterior myocardial infarction (AMI) is done using a simple adaptive threshold (SAT) method. The sensitivity, specificity and accuracy is 93.22%, 94.28% and 93.61% respectively in case of IMI and in AMI cases its 83.33%, 88.57%, 86.15% respectively. Classification is also done using artificial neural network, but its performance is comparatively low. This may be due to inadequacy of data available. The PTB diagnostic ECG database is used for evaluation of the methods