Publication Details

Category Text Publication
Reference Category Journals
DOI 10.23919/CinC53138.2021.9662908
Title (Primary) Arrhythmia classification of reduced-lead electrocardiograms by scattering-recurrent networks
Author Warrick, P.A.; Lostanlen, V.; Eickenberg, M.; Homsi, M.N.; Rodríguez, A.C.; Andén, J.
Source Titel Computing in Cardiology
Year 2021
Department MOLSYB
Volume 48
Page From 1
Page To 4
Language englisch
Topic T9 Healthy Planet
Keywords electrocardiography; scattering transform; phase harmonic correlation; canonical correlation analysis; convolutional neural networks; long short-term memory networks
Abstract We describe an automatic classifier of arrythmias based on 12-lead and reduced-lead electrocardiograms. Our classifier composes the scattering transform (ST) and a long short-term memory (LSTM) network. It is trained on PhysioNet/Computing in Cardiology Challenge 2021 data. The ST captures short-term temporal ECG modulations while reducing its sampling rate to a few samples per typical heart beat. We pass the output of the ST to a depthwise-separable convolution layer which combines lead responses separately for each ST coefficient and then combines resulting values across ST coefficients. At a deeper level, 2 LSTM layers integrate local variations of the input over long time scales. We train in an end-to-end fashion as a multilabel classification problem with a normal and 25 arrhythmia classes. We used canonical correlation analysis (CCA) for transfer learning from 12-lead ST representations to reduced-lead ones. For 12-, 6-, 4-, 3- and 2-leads, team “BitScattered” Challenge metrics on the hidden validation set were 0.46, 0.44, 0.45, 0.46 and 0.43; and on the hidden test set were 0.10, 0.11, 0.10, 0.10 and 0.10, respectively, ranking 34 th on the hidden test set.
Persistent UFZ Identifier https://www.ufz.de/index.php?en=20939&ufzPublicationIdentifier=26617
Warrick, P.A., Lostanlen, V., Eickenberg, M., Homsi, M.N., Rodríguez, A.C., Andén, J. (2021):
Arrhythmia classification of reduced-lead electrocardiograms by scattering-recurrent networks
Computing in Cardiology 48 , 1 - 4 10.23919/CinC53138.2021.9662908