Title :
Morphology-based measurement of activation time in human atrial fibrillation
Author :
Sandrini, L. ; Faes, L. ; Ravelli, F. ; Antolini, R. ; Nollo, G.
Author_Institution :
Dept. of Phys., Trento Univ., Italy
Abstract :
The measurement of the activation time is crucial to allow the correct automatic analysis and classification of intracardiac electrograms recorded in the human atria during atrial fibrillation (AF). This study proposes a method which accounts for the morphology of bipolar signals. After ventricular artifact removal and activation wave recognition, the fiducial point of the activation wave was set at its local barycentre (LB). The method was tested on a set of 30 AF bipolar recordings of increasing complexity class; its performance was compared with that of the traditional methods of maximum peak (MP) or maximum slope (MS) estimation, taking the manual measurements performed by an expert cardiologist as a reference. While for signals with low complexity the agreement with manual analysis was high for all automatic methods, in presence of disorganized AF characterized by complex morphology LB estimation resulted more reliable than MP or MS calculation.
Keywords :
electrocardiography; mathematical morphology; medical signal processing; signal classification; activation time; activation wave; activation wave recognition; automatic analysis; automatic classification; bipolar signals; complexity class; expert cardiologist; fiducial point; human atrial fibrillation; intracardiac electrograms; local barycentre; manual measurements; maximum peak estimation; maximum slope estimation; morphology-based measurement; performance; ventricular artifact removal; Atrial fibrillation; Band pass filters; Cardiology; Electrodes; Humans; Morphology; Performance analysis; Performance evaluation; Signal analysis; Time measurement;
Conference_Titel :
Computers in Cardiology, 2002
Print_ISBN :
0-7803-7735-4
DOI :
10.1109/CIC.2002.1166842