SEOUL — South Korean digital health company Huinno has achieved a major milestone in heart health technology, with a new study showcasing its artificial intelligence-powered arrhythmia detection model, ECG-GraphNet. The study was recently published in the international cardiology journal, Heart Rhythm O2.
The research, titled “ECG-GraphNet: Advanced Arrhythmia Classification Based on Graph Convolutional Networks,” highlights a groundbreaking method of analyzing electrocardiogram (ECG) signals using Graph Convolutional Networks (GCNs). By converting ECG data into graph structures, the AI model is able to more accurately classify different types of irregular heartbeats, or arrhythmias.
Unlike traditional time-series models, ECG-GraphNet transforms the electrical activity of the heart — captured through P-QRS-T waveforms — into graph formats. This allows the AI to learn both the timing and shape of the signals at the same time, improving diagnostic accuracy.
The model focuses on identifying three key types of heart rhythms:
- Normal (N)
- Supraventricular ectopic beat (S)
- Ventricular ectopic beat (V)
Performance was evaluated using the F1 score, a common metric that balances precision and recall, and the results indicated high classification accuracy.
A major advantage of ECG-GraphNet is its ability to process ECG data of varying lengths, which is essential for long-term or continuous monitoring. This makes the system particularly effective for use in wearable health devices.
Huinno plans to integrate ECG-GraphNet into its MEMO Cue system, a wearable solution designed for real-time heart monitoring. MEMO Cue consists of three parts:
- MEMO Patch, a wearable ECG sensor
- MEMO Link, a transmitter that sends data to hospital servers
- A centralized monitoring platform, where doctors can observe patients remotely
In March, MEMO Cue received reimbursement approval from South Korea’s Health Insurance Review and Assessment Service (HIRA) under the remote electrocardiographic surveillance code (EX871). This allows patients using the system to benefit from national insurance coverage.
The company sees this approval as a stepping stone to wider adoption. Huinno is now working to distribute MEMO Cue to hospitals and healthcare centers across the country.
“This publication proves the strength of our AI-driven ECG analysis and reflects our commitment to global research standards,” said Huinno CEO Gil Yeong-joon. “We will continue to expand our technology in the fields of wearable patient monitoring and diagnostic support.”
Chief Strategy Officer Kim added that the company aims to strengthen its presence in both domestic and global digital healthcare markets by pushing forward on both technology and commercialization.
Related topics:
