Hai Duong Nguyen, Xuan-The Tran
ECG-RAMBA is a new framework that improves ECG classification generalization by disentangling morphology and rhythm, achieving strong performance across different datasets without test-time adaptation.
The ECG-RAMBA framework is designed to improve the reliability of ECG (electrocardiogram) classification across different settings, which is crucial for clinical use and long-term monitoring. Traditional models often mix up waveform patterns and rhythm dynamics, leading to poor performance when applied to new datasets. ECG-RAMBA addresses this by separating these components and then combining them in a way that enhances generalization. The framework has shown promising results, particularly in detecting atrial fibrillation, and outperforms existing models in cross-dataset evaluations.
