Integrated shape averaging of the P-wave applied to AF risk detection
Résumé
The recognition of patients with atrial fibrillation (AF) risk, from body surface ECG, remains an actual problem. The aim of the work is to investigate how the study of the P-wave morphology may assess any clustering algorithm in separating patients with AF risk from healthy subjects. The clustering is done by the k-means algorithm and using a new method: the integrated shape averaging (ISA). According to the results, the method was able to separate patients with AF risk from safety ones with 75 % sensitivity and 70 % specificity on 40 recordings. Furthermore, two families of P-wave patterns were characterized among the patients with AF risk and probably corresponding to electrical manifestations of conduction defects in the left or the right atrium.