University of Hong Kong / Imperial College London
Model weights not public. Contact creators for more information.
CLIP-style model that aligns 12-lead ECG signals with free-text echocardiography reports for zero-shot detection of structural heart disease directly from an ECG. Extends the MERL framework, and was trained on 45,016 paired ECG-echo reports from two Hong Kong hospitals, with external validation on the public EchoNext dataset from Columbia University. Developed by researchers at the University of Hong Kong and Imperial College London; described in a 2025 medRxiv preprint.
Architecture
Hybrid
CLIP-style contrastive architecture aligning 12-lead ECG signals with free-text echocardiography reports, extending the MERL framework
Framework
PyTorch
Added to catalog
2026-07-10
100,000 de-identified 12-lead ECGs from 36,286 adult patients at Columbia University Irving Medical Center, each paired with echocardiogram-derived structural-heart-disease labels from a TTE within 1 year of the ECG.
LV systolic dysfunction (zero-shot)
Aortic stenosis (zero-shot)
Average across 6 SHDs (zero-shot)