Myocardial infarction detection and heart patient identity verification

This paper deals with a wavelet based method is used for detecting an Myocardial Infarction (MI) along with user identity. The multilayer Electrocardiogram signals is decomposed using Daubechies wavelet transform which segments the ECG into different sub bands. The inversion of the ST segment, QRS complex and PQ changes normally occur in a abnormal signal. It considerably varies the structure of multiscale matrices at different values with both energy and Eigen space values are observed. From that the mean and variance values are measured. Datasets, include both healthy control signal and various abnormal signals are used from the PhysikalischTechnischeBundesanstal(PTB)diagnosticn ECG database. SVM classifier is mainly used to classify the normal and abnormal cases in the signal. RR interval is used to authenticate the ECG signal. For MI detection the 90.42% accuracy is achieved for 15 ECG samples.