ECG classification for Sleep Apnea detection

Abstract

Sleep Apnea is a potentially serious sleep disorder in which you have one or more pauses in breathing or shallow breaths while you sleep. It is classified into 3 main types: Obstructive sleep apnea, Central sleep apnea, and Complex sleep apnea syndrome. Obstructive sleep apnea (OSA) represents 80% of the apnea cases which makes it the most common type. Polysomnography is the current traditional method used to diagnose OSA, it is expensive and needs human experts and done in a special laboratories, the need of a more comfortable and cheaper method arises recently to detect and diagnose such type of disorders. Recently researchers focused on signal processing and pattern recognition as alternative methods to detect OSA. In this paper, an automated classification algorithm is presented which processes short duration epochs of ECG data. The automated classification technique is based on three classifiers: Support vector machines (SVM), radial bases function (RBF), and multi-layer perception (MLP). The obtained results showed a high degree of accuracy, approximately 97.55 over passing all the other classifiers that have been already used in the literature. Moreover, the system we developed can be used as a basis for future development of a tool for OSA screening.