A semi-automatic QT interval measurement based on digital filters

The QT interval of ECG signal, a representative duration of cardiac cell activation, has been evidenced to correlate with cardiac disease or drug. Therefore the accurate QT measurement is important for clinical diagnosis and prognosis. In this paper we developed and presented a semi-automatic QT interval measurement method to participate the 2006 computers in cardiology challenge. This method locates the major markers of ECG signals (the P, Q, R, S T waves) based on digital filtering and predefined conditions in waveform. First a proper lead was select to measure the QT interval by manually inspection. The proper lead was defined as in which there are obvious upward R and T waves after 0.4-50 Hz band-pass filtering to diminish the noise and baseline wandering. Then R waves were found as local maxima after applying thresholding techniques. After the approximated first deviation of the filtered ECG (dECG) was obtained by difference filter, the Q onsets were located as the second cross-zero point of dECG 2-120 ms upstream R waves and the ends of T waves were located as the intercept of the filtered ECG signal and the line passing through T peak and maximum slope point on T wave. A visualization interface to display the characteristic markers on the ECG waveform was used to inspect the primary results and thus a representative beat could be manually selected as a final result. This method involves tow-steps manually intervention to raise the accuracy of QT measurement. However, these tow-steps are not time-consumed and easy for users with minimum knowledge of interpreting ECG. Finally an automatic method was developed by using the mean of 12 leads as processed signal.