Learning ECG Image Representations via Dual Physiological-Aware Alignments

2026-04-01

AI Summary

This study addresses the challenge that existing electrocardiogram (ECG) analysis methods rely on raw signals and struggle to leverage the vast archives of historical ECG data stored as images. To bridge this gap, the authors propose ECG-Scan, a novel framework that, for the first time, integrates image–signal–text multimodal contrastive alignment with soft lead constraints into self-supervised learning. By incorporating domain knowledge through a dual physiology-aware alignment mechanism, ECG-Scan enhances inter-lead consistency and improves clinical generalization during image representation learning. Experimental results demonstrate that the proposed method significantly outperforms current image-based baselines across multiple datasets and downstream tasks, substantially narrowing the performance gap between image-based and signal-based ECG analysis.

Abstract

Electrocardiograms (ECGs) are among the most widely used diagnostic tools for cardiovascular diseases, and a large amount of ECG data worldwide appears only in image form. However, most existing automated ECG analysis methods rely on access to raw signal recordings, limiting their applicability in real-world and resource-constrained settings. In this paper, we present ECG-Scan, a self-supervised framework for learning clinically generalized representations from ECG images through dual physiological-aware alignments: 1) Our approach optimizes image representation learning using multimodal contrastive alignment between image and gold-standard signal-text modalities. 2) We further integrate domain knowledge via soft-lead constraints, regularizing the reconstruction process and improving signal lead inter-consistency. Extensive benchmarking across multiple datasets and downstream tasks demonstrates that our image-based model achieves superior performance compared to existing image baselines and notably narrows the gap between ECG image and signal analysis. These results highlight the potential of self-supervised image modeling to unlock large-scale legacy ECG data and broaden access to automated cardiovascular diagnostics.

Problem

• ECG image
• raw signal recordings
• automated ECG analysis
• resource-constrained settings
• legacy ECG data

Innovation

• self-supervised learning
• multimodal contrastive alignment
• physiological-aware representation
• soft-lead constraints
• ECG image analysis