ImputeECG: Deep Learning Reconstruction of Complete 12-Lead Electrocardiograms from Incomplete Recordings for Cardiac Assessment
arXiv:2607.05009v1 Announce Type: cross Abstract: Complete digital 12-lead electrocardiograms (ECGs) are essential for AI-enabled cardiovascular assessment, yet many clinical ECG records, particularly those digitized from ECG images, remain incomplete because of short display formats, incomplete waveform digitization, lead loss, or signal corruption. We developed ImputeECG, a mask-conditioned one-dimensional Transformer autoencoder that completes 12-lead, 10-s ECGs while retaining all observed samples. The model was trained on PTB-XL and evaluated on PTB-XL and CPSC2018 under simulated incompl
Advances in transformer architectures and deep learning have reached a maturity that allows for their application to complex time-series reconstruction like medical signals.
Improving the completeness and utility of existing medical data, particularly ECGs, enables broader and more effective application of AI in cardiovascular diagnostics.
Previously incomplete or hard-to-digitize ECGs can now be converted into a standardized, AI-ready format, accelerating AI-driven medical research and diagnostics.
- · Healthcare AI developers
- · Cardiologists
- · Medical data platforms
- · Patients with cardiac conditions
- · ECG technicians relying on manual interpretation of poor quality data
- · Legacy ECG analysis software
Increased accuracy and accessibility of AI-powered cardiac diagnostics by addressing data quality issues.
Faster development and deployment of new AI models for early detection and personalized treatment of heart conditions.
Potential for remote and automated cardiac screening to become more widespread and reliable, reducing specialist burden and improving global access.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI