Challenge 2006: QT interval measurement

This years challenge addresses the question if it is possible to measure the QT interval by fully automated methods. For this task we implemented an algorithm which (1) performs beat detection, measures wave boundaries on a beat-to-beat basis and (2) selects a representative beat whose QT interval is used. To get the positions of the wave boundaries, the algorithm searches for peaks in the 1st derivative of bandpass filtered ECG signals using thresholds which are adapted to the amplitudes in the segments of interest. When the peaks associated with the QRS complex and the T wave are found, the beginning of the QRS complex and the end of the T wave are assessed using the "Philips Diagnostic ECG Interval Measurement" method [1]. For the selection of the representative beat, a noise level for each single beat was calculated by determining the standard deviation of the ECG 30ms before the onset of the QRS complex and 30ms after the end to the T wave. The beat with the lowest noise level was selected as the representative beat and from this the QT interval was used. This approach results in a score of 70.94ms.