COMBINING FORWARD-BACKWARD RR INTERVAL DETECTOR WITH ARRHYTHMIC BEAT CLASSIFICATION IN ECG SIGNALS

An electrocardiogram (ECG) provides valuable information about the functional aspects of the heart and cardiovascular system. This research develops a Forward-Backward (FB) algorithm to extract the QRS complex and combines with the arrhythmic beat classification rules on medical knowledge. The purpose of this study is to increase R-wave detection accuracy and monitor a patient's arrhythmia condition in real time. The FB algorithm augments forward and backward searching rules from QRS onset and eliminates lower-amplitude signals near the baseline, using a statistical process control (SPC) concept to improve R-wave detection performance. The MIT-BIH arrhythmia database (MITDB) is used as training data to determine the parameters of the proposed approach; it is classified into four arrhythmia beat types, i.e., normal (N), premature ventricular contraction (P VC), ventricular flutter/fibrillation (VF), or second-degree heart block (BII) beat. According to QRS complex extraction and beat classification rules, the proposed method provides 99.52% accuracy in the detection of R-wave location; 99.14% for N-classified beats; 93.37% for PVC-classified beats; 93.00% for VF-classified beats; and 89.852% for BII-classified beats in the MITDB conclusion, the proposed procedure could prove useful, because it uses only the RR interval for arrhythmic beat classification and is affected only slightly by noisy and complicated signals.