Aims: The ST segment change (elevation or depression) is a crucial symptom in the electrocardiogram (ECG) related with myocardial ischemia. Therefore, the detection of ST segment change plays an important role in the automatic ECG analysis algorithm. In this study, we proposed a new method to distinguish the ST segment change from the normal ST and classify the type of ST segment changes.
Methods: At First, the pre-processing and delineation of the fiducial points of ECG waves were applied to the ECG signal. Then, we extracted the features such as the time-frequency distribution, the complexity, the fractal dimension and the amplitude characteristics of the ST segment for each heartbeat. Finally, the support vector machine was used to recognize these features, and the type of heartbeat was classified into normal heartbeat, heartbeat with elevated ST segment or depressed ST segment.
Results: Nine records from the European ST-T database were used to evaluate the performance of the algorithm. The number of normal heartbeat, the heartbeat with ST segment elevation and depression was 42414, 3715 and 3007 respectively. For the depressed ST segment, the sensitivity and specificity of the classification were 95.2% and 75 respectively. And the sensitivity and specificity of the elevated ST segment were 92.5% and 87.9% respectively.
Conclusion: Experimental results shows that the developed algorithm was useful to automatically detect the ST segment elevation and depression, and it was meaningful to provide more details for the automatic analyze of the myocardial ischemic.