Efficient Early Detection of Cardiovascular Disease from ECG Imaging Using a One-Stage Deep Learning Model

Cardiovascular diseases (CVD) continue to rank among the world's top causes of death, highlighting the significance of early and precise diagnostic techniques. In this study, we propose deep learning-based approach that uses ECG imaging to diagnose cardiovascular illnesses early. The technique makes use of YOLOv8, a cutting-edge one-stage deep learning model renowned for its quick and precise object identification skills. Two well-known sources of the ECG dataset used in this work are the PTB Diagnostic ECG Database and the MIT-BIH Arrhythmia Dataset, which offers a thorough collection of heartbeat signals for categorization. The model was assessed using several key metrics, such as the precision-recall curve (0.92), recall confidence curve (0.99), F1 confidence curve (0.86), and precision confidence curve (0.99). These results demonstrate the model's excellent level of recall and precision., which makes it appropriate for accurate early CVD diagnosis. This study demonstrates the promise of YOLOv8 as a time-efficient solution with robust performance measures for ECG imaging-based diagnosis. Incorporating real-time diagnostic skills in clinical settings and improving model generalization across a variety of patient demographics will be the main goals of future research.