TY - CHAP
T1 - Deep Discriminative Domain Generalization with Adversarial Feature Learning for Classifying ECG Signals
AU - Shang, Zuogang
AU - Zhao, Zhibin
AU - Fang, Hui
AU - Relton, Samuel
AU - Murphy, Darcy
AU - Hancox, Zoe
AU - Yan, Ruqiang
AU - Wong, David
N1 - Publisher Copyright:
© 2021 Creative Commons.
PY - 2021
Y1 - 2021
N2 - Introduction: The goal of the 2021 PhysioNet/CinC challenge is to classify cardiac abnormalities from ECGs and evaluate the diagnostic potential of reduced-lead ECGs. Here, we describe the classification model created by the team 'AIHealthcare'. Methods: ECGs were downsampled to 300 Hz and filtered by wavelet. ECGs were randomly clipped or zero-padded to 4,096 samples. We modified an SE-ResNet to perform multi-task classification of both dataset and disease. We used a gradient reversal layer as part of an adversarial feature learning scheme to learn domain-invariant and discriminative representations. Results: We trained our domain-invariant model on 5 datasets, keeping one data set (Ningbo) for local validation. We also trained a baseline SE-ResNet using the same training data. In validation on the held-out data set, the domain-invariant model had a higher Challenge metric than the baseline model. Our entry was not officially ranked in the Challenge, as we did not have a successful entry during the unofficial phase of the Challenge. Conclusion: The domain-invariant model performed better than the baseline model in local held-out datasets, suggesting that this method may help improve generalisation performance.
AB - Introduction: The goal of the 2021 PhysioNet/CinC challenge is to classify cardiac abnormalities from ECGs and evaluate the diagnostic potential of reduced-lead ECGs. Here, we describe the classification model created by the team 'AIHealthcare'. Methods: ECGs were downsampled to 300 Hz and filtered by wavelet. ECGs were randomly clipped or zero-padded to 4,096 samples. We modified an SE-ResNet to perform multi-task classification of both dataset and disease. We used a gradient reversal layer as part of an adversarial feature learning scheme to learn domain-invariant and discriminative representations. Results: We trained our domain-invariant model on 5 datasets, keeping one data set (Ningbo) for local validation. We also trained a baseline SE-ResNet using the same training data. In validation on the held-out data set, the domain-invariant model had a higher Challenge metric than the baseline model. Our entry was not officially ranked in the Challenge, as we did not have a successful entry during the unofficial phase of the Challenge. Conclusion: The domain-invariant model performed better than the baseline model in local held-out datasets, suggesting that this method may help improve generalisation performance.
UR - http://www.scopus.com/inward/record.url?scp=85121297621&partnerID=8YFLogxK
UR - https://www.mendeley.com/catalogue/25dea0bd-9e25-3cb1-94e5-d924f8fcb4e6/
U2 - 10.23919/CinC53138.2021.9662844
DO - 10.23919/CinC53138.2021.9662844
M3 - Chapter
SN - 9781665479165
T3 - Computing in Cardiology
BT - Computing in Cardiology
T2 - 2021 Computing in Cardiology (CinC)
Y2 - 13 September 2021 through 15 September 2021
ER -