DocumentCode :
3562134
Title :
Estimation of atrial fibrillatory frequency by spectral subtraction of wavelet denoised ECG in patients with atrial fibrillation
Author :
Goodfellow, Jonathan ; Escalona, Omar J. ; Walsh, Philip R. ; Kodoth, Vivek ; Manoharan, Ganesh
Author_Institution :
Univ. of Ulster, Newtownabbey, UK
fYear :
2014
Firstpage :
329
Lastpage :
332
Abstract :
The objective of this study was to assess the efficacy of a novel wavelet based dominant atrial fibrillatory frequency (DAFF) estimation technique on twenty patients who underwent internal cardioversion at Royal Victoria Hospital, Belfast. The results acquired using the wavelet technique were compared against results achieved using a conventional average template subtraction (ATS) method with three performance parameters: percentage noise reduction ratio, ventricular activity (QRST) correlation value and percentage DAFF similarity ratio. The results for wavelet based estimation were 87.8%±5.3% and 0.983±0.006 respectively compared to 72.9%±9.3% and 0.981±0.010 for the ATS method. These results indicate that the wavelet technique offers improved performance at attenuating noise in the Q-T intervals than conventional technique, yet there remains high retention of the ventricular morphology, suggesting that the proposed wavelet technique could be a useful method for isolating atrial activity for DAFF analysis. DAFF estimated values using both techniques were compared and found to be in good agreement, as percentage DAFF similarity ratio was 94.8%±3.8%.
Keywords :
cardiovascular system; correlation methods; electrocardiography; medical signal processing; signal denoising; wavelet transforms; ATS; DAFF analysis; DAFF estimated values; Q-T intervals; QRST; Royal Victoria Hospital; atrial activity; conventional average template subtraction method; internal cardioversion; percentage DAFF similarity ratio; percentage noise reduction ratio; spectral subtraction; ventricular activity correlation value; ventricular morphology; wavelet based dominant atrial fibrillatory frequency estimation technique; wavelet based estimation; wavelet denoised ECG; Abstracts; Estimation; Kalman filters; Morphology;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computing in Cardiology Conference (CinC), 2014
ISSN :
2325-8861
Print_ISBN :
978-1-4799-4346-3
Type :
conf
Filename :
7043046
Link To Document :
بازگشت