Echo2ECG: Enhancing ECG Representations with Cardiac Morphology from Multi-View Echos

Echo2ECG is a multimodal self-supervised learning framework designed to enhance ECG representations by integrating cardiac morphology from multi-view echocardiograms. It serves as an ECG feature extractor for two key tasks: classifying structural cardiac phenotypes and retrieving echocardiographic studies based on ECG queries. The model, significantly smaller than existing baselines (18x smaller), demonstrates superior performance in both tasks, indicating its potential for improving early health screening and diagnostics in cardiology.