An end-to-end deep learning pipeline for automated ECG rhythm classification using ambulatory electrocardiogram recordings. This project focuses on distinguishing between Normal Sinus Rhythm (NSR), Atrial Fibrillation (AF), Other Arrhythmias, and Noisy Recordings from single-lead ECG signals. The model was trained on 8,500+ labeled ECG recordings ranging from 9 to 61 seconds in duration, with preprocessing, feature extraction, model training, and evaluation integrated into a unified workflow.
