Idea
Dataset powering AI models for accurate, scalable ECG diagnosis across diverse clinical settings.
Research Paper
Core Innovation
This paper introduces CODE-II, a large-scale, real-world ECG dataset with 66 clinically meaningful diagnostic classes annotated and reviewed by cardiologists. It surpasses prior datasets in size, annotation quality, and clinical relevance, enabling superior AI model training and transfer performance on external benchmarks.
Why It Matters
Accurate ECG interpretation is critical for timely cardiac care but is limited by data quality and scale. CODE-II addresses these gaps by providing a large, well-annotated dataset that improves AI diagnostic accuracy and generalizability. This enables healthcare providers to scale ECG analysis efficiently and improve patient outcomes globally.
Market Size (TAM)
$10–20B TAM for AI-driven cardiac diagnostics; $2–5B SAM from hospitals, telehealth, and medical device sectors. Driven by rising cardiovascular disease prevalence and demand for scalable diagnostic tools.
Potential Customers & Pain Points
- Hospitals – Need reliable AI for ECG diagnosis
- Telehealth providers – Require scalable ECG interpretation
- Medical device companies – Seek robust training data for AI algorithms
- Research institutions – Need high-quality ECG datasets for model development
Business Model
Licensing the dataset and pretrained models to healthcare providers, medical device manufacturers, and research institutions; offering API access for AI model development and validation.
Competitive Landscape
- PTB-XL dataset
- CPSC 2018 dataset
- PhysioNet ECG datasets
Implementation Challenges
- Regulatory approval for AI diagnostic tools
- Integration with existing clinical workflows
- Data privacy and security concerns
Validation Strategy
- Conduct blinded clinical trials comparing AI models trained on CODE-II versus existing datasets
- Partner with hospitals and telehealth networks for real-world deployment and feedback
- Benchmark model performance on external ECG datasets and regulatory standards
Research Paper Overview
CODE-II: A large-scale dataset for artificial intelligence in ECG analysis
Summary
CODE-II is a comprehensive dataset of over 2.7 million 12-lead ECGs from more than 2 million patients, annotated with 66 clinically relevant diagnostic classes and reviewed by cardiologists. It includes a public subset and a test set for blinded evaluation. Models pre-trained on CODE-II show superior transfer learning performance on external ECG benchmarks.