New method for QRS-wave recognition in ECG using MART neural network

Abstract

Recognition of QRS-Wave in the ECG signal is one of the important stages for ECG signal processing and most of the ECG noise removal algorithms, and automatic ECG interpreter systems need to detect these points. In most cases ECG signals are noisy and we need to detect these points using noisy signals. We have developed a QRS-wave recognition system using MART (multi-channel ART) neural network. In this method signal of two leads of ECG is used for detection, so our method has low sensitivity to noises. We tested our method for noiseless and noisy ECG signals and we compared results against those of an older one, which uses ART2 neural network. Results showed that our method has good results for noisy signals.