A ROBUST AND EFFICIENT ALGORITHM FOR ST–T COMPLEX DETECTION IN ELECTROCARDIOGRAMS

Abstract

Changes of ST–T complex are the main indicators in myocardial ischemia detection based on electrocardiogram (ECG) signals. However, ST–T complex is highly sensitive to interferences (baseline wandering, postural changes, electrode interference, etc.), especially T-wave has a large morphological variability. Therefore, the feature points of ECG ST–T complex are very difficult to detect accurately. Currently, the commonly used detection methods for ST-segment include R + x and J + x, but they often misjudges the T-wave rising limb as ST-segment. Therefore, a new accurate hybrid approach for ST–T detection was proposed in this paper. First, T-wave onsets and offsets were detected using regional method and T-wave peaks were located using function comparing method. Then, a squeeze approach for ST-segment detection was proposed based on R-wave peak and T-wave onset. Long-term ST database (LTST) verification demonstrated that the accuracy of ST–T detection reached above 91%. In addition, it had a good timeliness and robustness and was easy to implement.