Key result
An EfficientNet-B3 deep learning model achieved challenge evaluation metric scores of 0.45 to 0.48 for identifying clinical diagnoses from reduced-lead and 12-lead ECG recordings.
Why the study?
The study was conducted to identify clinical diagnoses from 12-lead and reduced-lead ECG recordings as part of the PhysioNet/Computing in Cardiology Challenge 2021.
An EfficientNet-B3 deep learning model demonstrated competitive performance in identifying clinical diagnoses from both standard 12-lead and reduced-lead ECG recordings.
Requires further validation before clinical ECG use; leaves open avenues for refining self-supervised models on reduced-lead data.
The goal of PhysioNet/Computing in Cardiology Challenge 2021 was to identify clinical diagnoses from 12 -lead and reduced-lead ECG recordings, including 6-lead, 4-lead, 3-lead, and 2-lead recordings. Our team, snu_adsl, have used EfficientNet-B3 as the base deep learning model and have investigated methods including data augmentation, self-supervised learning as pre-training, label masking that deals with multiple data sources, threshold optimization, and feature extraction. Self-supervised learning showed promising results when the size of labeled dataset was limited, but the competition's dataset turned out to be large enough that the actual gain was marginal. In consequence, we did not include self-supervised pre-training in our final entry. Our classifiers received scores of 0.48, 0.48, 0.47, 0.47, and 0.45 (ranked 12th, 10th, 11th, 11th, and 13th out of 39 teams) for the 12-lead, 6-lead, 4-lead, 3-lead, and 2 -lead versions of the hidden test set with the Challenge evaluation metric.
No takes yet. Share an insight, caveat, or question.
Suh et al. (2021) studied Cardiac abnormalities. EfficientNet-B3 deep learning model was evaluated on Challenge evaluation metric score for 12-lead, 6-lead, 4-lead, 3-lead, and 2-lead ECGs. An EfficientNet-B3 deep learning model achieved challenge evaluation metric scores of 0.45 to 0.48 for identifying clinical diagnoses from reduced-lead and 12-lead ECG recordings.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: