Multilead QT Interval Analysis Algorithm Based on Continuous Wavelet Transform

İŞCAN M., Sariozkan S., Yilmaz A. , YILMAZ C.

4th International Conference on Electrical and Electronic Engineering (ICEEE), Ankara, Türkiye, 8 - 10 Nisan 2017, ss.318-322 identifier identifier


Computer-aided recognition program is being established to detect irregular patterns of cardiac rhythm causing several mortality conditions such as heart failure, heart attack, etc. In this study, an automatic novel algorithm based on continuous wavelet transform (CWT) conducting multi-lead process was proposed for estimating QT interval in ECG signals. Especially, T wave end delineation is the major problem in ECG processing. In the first place, the presented algorithm was utilized to locate the starting point of QRS complexes and the final point of T wave. The suggested algorithm was assessed on QT database published by PHYSIONET, then, the achievements of the algorithm were mean error of -1.2423 and -0.2136 milliseconds with the standard deviation error of 18.3490 and 7.6522 milliseconds at silver and gold standard, respectively. As a result, the suggested algorithm is a fully automatic method with high performance characteristic in multi-lead ECG analysis.