HBTrackr: AI-Based Heart Attack Prediction through ECG Monitoring on Wearable Devices

The topic proposes the existence of various mobile applications designed to predict and detect abnormal heartbeats, providing individuals with an accessible tool for monitoring their cardiac health. Advancement in AI supports heart attack prediction which can be deployed in various smart applications. In the paper, a novel AI model is implemented to read the Electrocardiogram (ECG) of the patient and understand the heart health of the person by wearing a smart watch. The proposed framework focuses on detection of sudden cardiac arrest and prediction of abnormal heartbeat using You Only Look Once(YOLOv8) in the wearable devices. The model is trained on ECG images dataset of cardiac patients using a matrix of features such as myocardial infarction, irregular heart rhythms, and healthy cardiac patterns. The model achieves high precision, recall, and F1 score, demonstrating its effectiveness in accurately classifying cardiac conditions. HBTrackr is an easy-to-use smartwatch app which is incorporated with an AI model that uses the pre-existing sensors in the watch.