QT Interval Measurement based on Singular Value Decomposition

This paper describes a method for accurate and stable measurement of ECG QT intervals. ECG signals from multiple leads are decomposed into orthogonal signals by the algebraic method of singular value decomposition. Major orthogonal signal components are then selected and root total power signal is formed. The root total power signal is found to be stable in its waveform and suitable for Q and T-end time determination by the aid of either least square or tangential line fit. Illustrative examples are chosen from normal, long QT syndrome (LQTS) and left ventricular hypertrophy (LVH) subjects to demonstrate the effectiveness of the proposed method. Automatically measured QT intervals are also compared to those estimated by eye fit of an experienced cardiologist