Review of open foundation models and datasets for ECG and PPG waveforms

Le résumé fourni par la source

Foundation models (FMs) offer a promising approach for electrocardiography (ECG) and photoplethysmography (PPG) analysis, enabling reusable representations across tasks without large labeled datasets or task-specific training. We review open ECG/PPG FMs and datasets, outlining cardiovascular applications. Advantages include data efficiency and generalizability, though challenges remain, particularly reliance on homogeneous pretraining datasets and the absence of standardized evaluation frameworks. This review guides clinicians in understanding and applying FMs in cardiovascular research.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.