Skip to Main Content
A new method is presented for detection and classification of arrhythmias in electrocardiograms on the basis of R-R interval data. A set of phenomenological models for both persistent and transient rhythms is developed to match observed statistical variations. Arrhythmias are identified by calculating statistical probabilities and likelihoods associated with these models using two recently developed techniques. The important system design considerations are described. Finally, representative results using actual arrhythmia data are presented to illustrate the system performance.