AI-Powered Silent AFib Detection and Ischemic Stroke Risk Prediction via 3-Lead ECG
CWSF · 2026 Disease & Illness Silver Medal
Overview
Every 40 seconds someone has a stroke and up to 30% are caused by a heart condition called atrial fibrillation that produces zero warning signs until it's too late. We built a portable heart monitor using a $12 Arduino device with three chest electrodes that records 60 seconds of heart activity and feeds it into a deep learning model we trained on over 8,500 real patient recordings. Our model correctly identified 103 out of 111 real atrial fibrillation cases in patients it had never seen before, achieving accuracy that matches cardiologist-level detection. It then calculates a personalized annual stroke risk percentage and shows how much that risk drops if the patient starts treatment. A hospital heart monitor costs $800; ours costs $12, requires no specialist, and could detect this silent condition years before a stroke ever happens.
Awards (2)
- Silver Medal
- Selected for CWSF 2026
Competition history
- CWSF 2026
Related projects
ISEF · 2025
RhythmTrack AI- A Cost-Effective Stationary Monitoring Device for the Detection of Atrial Fibrillation (AF) Using a Three-Electrodes Configuration on Einthoven's Triangle by CNN-LSTM-Based AI Model
ISEF · 2018
The Development of a Holistic Cardiovascular Disease Screening System Utilizing a Low-Cost Electrocardiogram and a Machine Learning Algorithm
ISEF · 2020
An Active Role for Machine Learning in the Diagnosis of Atrial Fibrillation
ISEF · 2026
GuardianLink: A Personalized Neuro-Cardiological Health-Monitoring System for Real-Time Heart Arrhythmia Detection, Prediction, and Analysis Through an Adaptive Machine Learning Framework With an ECG Necklace and Biometric Smartwatch
Closest projects by meaning, across every fair and year in the corpus.