Automatic detection of coronary events by wavelet transform and delineation of fiducial points on ECG signals
AbstractIn the present work the wavelet transform was used for the detection of the R peaks of the electrocardiogram (ECG) signal and a procedure was developed to delineate the other fiducial points of said signal based on search algorithms of maximum, minimum, zero crossing points and inflection points, all of them with low computational cost, in order to facilitate the automatic detection of coronary events by measuring the ST segment and T wave deviations. The algorithm was implemented, with interactive interface, using the Matlab programming language. Validation tests with both normal and pathological cardiac signals were performed using European ST-T ischemic event database records available on the Physionet portal. Sensitivity and specificity were calculated, obtaining as results 97.5% and 94.3%, respectively, for the ST segment deviation and 95.7% and 92.2%, respectively, for the T wave deviation. The algorithm developed has as its main advantages its high processing speed and high effectiveness, necessary characteristics for its implementation in portable systems.