Exploiting Cross-Correlation Between ECG signals to Detect Myocardial Infarction

Electrocardiograph (ECG) signals are used to monitor heart activities, which is a non-invasive technique. To detect a problem in the heart, an expert is required to read these signals and diagnose a disease. In this paper, a simple and fast automatic technique is proposed to know if a person is going through a Myocardial Infarction. The proposed technique exploits the cross-correlation between the ECG signals and does not require an expert to read ECG signals. Data acquired from an online resource contains baseline wanders and noise. To remove this noise discrete-wavelet-transform and moving average filters are used. A threshold of 0.54 is chosen for the classification of the patient as a normal or going through myocardial infarction. If the cross-correlation is above 0.54, the patient is classified as normal and vice versa. Our simulation results show that the accuracy, specificity, and sensitivity of the proposed technique are respectively 91.5%, 86.5%, and 95.6%.