In this paper, the performance-three QRS cancellation techniques compared to obtain a residual ECG information containing only atrial activity (AA), thereby facilitating detection of the P wave in the ECG and/or analysis of atrial arrhythmias. The methods are based on adaptive filters (AdF), neural networks (NN), and the space-time of an averaged beat for subsequent removal from the ECG signal, the comparison is made with reference to the results obtained with the adaptive filtering method by Thakor. The results show that the methods based on NN and AdF are more robust to morphological changes of the signal and in the presence of noise than non-adaptive methods, also have the advantage that they do not require a pre-detection stage of QRST, however, for the average beat alignment method, best results are obtained when there are no QRS detection failures.
|Journal||Technology and Health Care|
|Publication status||Published (in print/issue) - 15 May 2001|
- QRS cancelation techniques
- atrial fibrillation ECG
- adaptive filters
- neural networks.