Individually adaptable automatic QT detector

The 7th Computers in Cardiology challenge presents the clinically important problem of developing and testing fully automated methods for measurement of the QT interval with clinically acceptable accuracy. For this purpose a fully automated method was developed using time-frequency methods of continuous (CWT) and fast wavelet transforms (FWT). It consists of QRS detection and P, T wave boundaries identification which is implemented using consecutive ECG data filtering with CWT and FWT transforms. The frequency of the CWT transform can be adjusted for more precise identification of the T wave boundaries for an individual patient. A representative beat with a T wave is selected from an automatically annotated signal if it is not annotated as ectopic, noise or artefact. The final score is the highest one in division three and second place in division two.