DATABASE FOR VALIDATION OF METHODS OF EXTRACTION OF ATRIAL FIBRILATORY WAVES IN THE ECG
This is a database composed of simulated ECG signals for validation of f-wave extraction algorithms. This dataset not only facilities the comparison of performance, but it highlights a broader set of signal characteristics than the dates which so far have been used in previous studies. The included signals features are well-known to represent major challenges in f-wave extraction, therefore including high heart rates, high morphological QRST variability, the presence of ventricular premature beats, different f-wave amplitude and repetition rates, and f-wave morphology involving a stochastic component.
The database includes a total of 240 signals, which are divided into eight sets (labeled S1-S8). Each set includes 30 signals with six different values of amplitude (5, 10, 20, 30, 40, 50 uV), 5 minutes in length and different combinations of the above-mentioned features. These sets of signals can be downloaded from the links below.
For more details, please refer to the work: R. Alcaraz, JJ. Rieta, L. Sörnmo, «Reference Database and Performance Evaluation of Methods for Extraction of Atrial Fibrillatory Waves in the ECG», Physiological Measurement, 2019: under review.