APPLICATIONS OF ARTIFICIAL NEURAL NETWORKS FOR ECG SIGNAL-DETECTION AND CLASSIFICATION

The authors have investigated potential applications of artificial neu ral networks for electrocardiographic QRS detection and beat classific ation. For the task of QRS detection, the authors used an adaptive mul tilayer perception structure to model the nonlinear background noise s o as to enhance the QRS complex. This provided more reliable detection of QRS complexes even in a noisy environment. For electrochardiograph ic QRS complex pattern classification, an artificial neural network ad aptive multilayer perception was used as a pattern classifier to disti nguish between normal and abnormal beat patterns, as well as to classi fy 12 different abnormal beat morphologies. Preliminary results using the MIT/BIH (Massachusetts Institute of Technology/Beth Israel Hospita l, Cambridge, MA) arrhythmia database are encouraging.