Rhythm Analysis by Heartbeat Classification in the Electrocardiogram

摘要

The morphological and rhythm analysis of the electrocardiogram (ECG) is based on ventricular beats detection, wave parameters measurement, as amplitudes, widths, polarities, intervals and relations between them, and a subsequent classification supporting the diagnostic process. Number of algorithms for detection and classification of the QRS complexes have been developed by researchers in the Centre of Biomedical Engineering ? Bulgarian Academy of Sciences, and are reviewed in this material. Combined criteria have been introduced dealing with the QRS areas and amplitudes, the waveshapes evaluated by steep slopes and sharp peaks, vectorcardiographic (VCG) loop descriptors, RR intervals irregularities. Algorithms have been designed for application on a single ECG lead, a synthesized lead derived by multichannel synchronous recordings, or simultaneous multilead analysis. Some approaches are based on templates matching, cross-correlation or rely on a continuous updating of adaptive thresholds. Various beat classification methods have been designed involving discriminant analysis, the K-th nearest neighbors, fuzzy sets, genetic algorithms, neural networks, etc. The efficiency of the developed methods has been assessed using internationally recognized arrhythmia ECG databases with annotated beats and rhythm disturbances. In general, high values for specificity and sensitivity competitive to those reported in the literature have been achieved