Automatic Detection Of QT Interval In ECG Signals

The study of ECG wave characteristics has always been one of the research focuses in biomedical engineering. The absolute prolongation in the QT interval is of great importance to the prognosis of acute miocardial infarction, the diagnosis of cardiac muscle damage, some drug toxicosises, electrolyte turbulence and long QT interval syndrome (LQTS). Especially after the definitude of the relation between the prolongation in the QT interval and the malignant arrhythmia, some problems involed with the measurement and analysis of QT interval have attracted new clinical attentions.In this paper, firstly, we used the structuring elements pair consisted of a triangular structuring element and a short line segment structuring element as morphological filter which could not only reduce the small fluctuations but also save the global fluctuating trend. This design avoided the ladderlike waveform appearing in the signal filtered by single short line segment structuring element.Secondly, according to the different lengths of QRS waves, different scales were set to do the multiscale morphological derivative (MMD) transform in different wave. According to our experiments, we found the condition of wave classification and the appropriate value of threshold and transform scales. When these were all settled down, this detector could process lots of signals with various waveforms in real time system.Thirdly, by introducing the'wing'function, two referenced points of T peak were found. The variety of T wave was defined according to the amplitudes of these two referenced points, which was helpful to locate the T peak point exactly. This improved the performance of MMD detector in detecting the bifid T wave and biphasic T wave. Doing the MMD transform with adaptive scale based on the different characteristics could hold and amplify the characteristic wave more efficiently than that with the single scale. Moreover, it distinguished between the positive T wave and negative T wave by adding an estimator which got better results than the former method in the detection of negative T wave and biphasic T wave.Finally, by testing the method with the records from CSE database, the MMD showed great superiority and potential in automatic measurement of QT interval.