Effective Wavelet Algorithm for an Automatic Detection of Atrial Fibrillation

Abstract

Cardiac anomalies are usually marked through irregular cardiac cycles. Atrial fibrillation is given through a rapid beating of the atria, announcing a possible heart failure or stroke. Electrocardiograms are an efficient way of supervising the electric activity of the heart. We have developed an effective, simple to implement automatic detection algorithm for identifying changes of the cardiac rhythm. The algorithm is based on wavelets and an enhanced time domain thresholding procedure. We take into account a variation of the electrocardiogram’s amplitudes, to avoid loss of clinical features. The interval between beats is computed and provided for a reliable diagnosis. The results are validated both with objective evaluation criteria and displayed graphically, assisting the medical diagnosis procedure.

Keywords

• adaptive algorithms
• biomedical signal processing
• performance evaluation
• threshold
• wavelet