Advancing ECG Signal Analysis with a Novel Geometric Feature Space for Differentiating Atrial Fibrillation

Análise de ECG por Meio de um Novo Espaço de Características Geométricas para Detecção de Fibrilação Atrial

Abstract

Electrocardiogram (ECG) signals are inherently challenging to classify due to their complex and variable nature. This study introduces a novel approach for classifying normal and atrial ECG signals by utilizing a geometric feature space (GFS) generated from three specific geometric characteristics: amplitude, Zenith angle, and shape factor. By employing a sweeping technique to extract and compile unique information, the proposed algorithm constructs this feature space, which is subsequently processed by standard machine learning models to classify the two signal sets. Our methodology demonstrates substantial improvements in classification performance, achieving a maximum F1-score of 0.99 %. These results underscore the effectiveness and robustness of our approach in distinguishing between normal and atrial ECG signals.