An intelligent sleep apnea detection system

This paper proposes an intelligent sleep apnea detection system. The system utilizes a band-pass filter to filter out components with extreme low and high frequencies from the electroencephalogram. It only conserves sleep-related bands from 0.5 to 32Hz. Moreover, it extracts frequency elements from Hilbert spectrum by Hilbert-Huang transformation. The system computes four Hilbert spectrum frequency intensities of the alpha, beta, theta, and delta signal from the transformed spectrum. The system then detects duration of obstructive sleep apnea from the frequency variation. The experimental results show that system can detect the duration of OSA as well as preserve time information in the electroencephalogram by Hilbert-Huang transformation mechanism and find frequency variation information.