Development of FCM based algorithm for the delineation of QRS-complexes in Electrocardiogram

This paper proposes a simple technique for automatic detection and delineation of cardiac beat (QRS-complex) in Electrocardiogram (ECG) using Fuzzy C-Means (FCM) clustering algorithm. The power line interference and baseline wander present in the ECG signal is removed using digital filtering techniques. Absolute slope of the filtered ECG signal is calculated to enhance the QRS-complexes in the ECG signal. FCM technique is then used to delineate QRS and non-QRS-regions. Algorithm performance is validated using original single lead ECG recordings from the dataset-3 of the CSE multi-lead measurement library. Detection rate of 98.74% is achieved. The percentage of false positive and false negative detection is very low. The delineation performance of the algorithm is validated by calculating the mean and standard deviation of the errors between automatic and manual annotations by the referee´s cardiologists/combined program median in the CSE study.