Automatic Detection of Myocardial Infarction Based on High-Frequency QRS Analysis

In this paper, we present an algorithm for automatic detection of myocardial infarction using high-frequency components of the ECG signal. Firstly, the QRS complexes and their boundaries are identified. Then, the correlation matrix between the detected QRS complexes in each lead is determined to eliminate noises and ectopic oscillations. The dominant QRS complexes are finally determined using cluster analysis. These resulting values are averaged to have a unique representative QRS complex in a given lead. This averaged signal is then passed through a band-pass filter to obtain high-frequency components of the QRS complex. Finally, the High-Frequency Morphological Index (HFMI) for each lead is calculated and diagnosed with myocardial infarction based on decision rules. The performance of the proposed algorithm is evaluated on signals from the PTB database. The obtained results show that the proposed method reached satisfactory performance compared with the results from clinical studies. Keywords: Myocardial infarction, High-frequency ECG, RAZ, RMS, HFMI.