Convolution-Free Waveform Transformers for Multi-Lead ECG Classification
This work addresses ECG classification for medical diagnosis, but it is incremental as it applies a transformer model to an existing challenge without major methodological breakthroughs.
The paper tackled detecting cardiac abnormalities from ECG recordings using a waveform transformer model, achieving an average challenge metric of 0.47 on a held-out test set across various ECG-lead subsets and ranking 11th out of 39 teams in the PhysioNet/CinC challenge.
We present our entry to the 2021 PhysioNet/CinC challenge - a waveform transformer model to detect cardiac abnormalities from ECG recordings. We compare the performance of the waveform transformer model on different ECG-lead subsets using approximately 88,000 ECG recordings from six datasets. In the official rankings, team prna ranked between 9 and 15 on 12, 6, 4, 3 and 2-lead sets respectively. Our waveform transformer model achieved an average challenge metric of 0.47 on the held-out test set across all ECG-lead subsets. Our combined performance across all leads placed us at rank 11 out of 39 officially ranking teams.