ECG ST segment change detection using Born-Jordan Time-Frequency Transform and Artificial Neural Networks

Early detecetion of ST segment's depression or elevation in ECG signal very important for prevention of heart attack and correct diagnosis and treatment after a heart attack. In this study, an algorithm based on Born-Jordan Time Frequency Transform was developed in order to early detection of ST segment's depressions or elevations. The performance evaluation of the algorithm was performed on a large database produced from MIT-BIH arrhythmia and European ST-T databases. From the MIT-BIH database, 111688 R-R intervals containing healthy or arrhythmias in V1, V2, V4, V5 leads and R-R intervals with 111688 ST segment's depression or elevation in V1, V2, V3, V4, V5 leads from the European ST-T database were selected. Artificial Neural Networks method was used in the classification stage. The classification performance results were found as accuracy, sensitivity, specificity, positive predictive value, negative predictive value and F score of 98.20%, 97.93%, 98.52%, 98.46%, 97.95% and 98.19% respectively.