Spatio Temporal Filtering of Multi-lead ECG Signals for Atrial Arrhythmia Classification

Abstract

Atrial fibrillation (AF) is the most common cardiac arrhythmia and increases the risk of suffering stroke. Some people with AF do not have symptoms, so, its diagnosis can be difficult, especially in early stages of the disease. In this paper, we propose the use of the spatio-Temporal filter (STF) to characterize atrial activity in ECG recordings and distinguish between normal sinus rhythm (NSR) and atrial arrhythmias. This method allows the effective detection of P waves when they are synchronized with QRS complexes. The distances from the QRS complexes to the detected P waves are characterized by seven dispersion metrics that are used as inputs to three clustering algorithms. The results show classification accuracy of up to 98.88% of NSR and atrial arrhythmias.