Classification of Oral/Nasal Simulated Snores Based on the Acoustic Properties

Abstract

Snoring was once regarded as an indication of good sleep, but recently it has been known to be one of the symptoms which indicate sleep disordered breathing such as sleep apnea syndrome. Especially, loud snoring caused by oral breathing during sleep is often found in many apnea/hypopnea patients. Thus, it is important to detect oral snoring for medical treatment in the earlier stage, but we cannot know our own snoring. This paper describes a method to detect oral snoring by extracting the acoustic properties of snoring sounds. According to the FFT amplitude spectra, nasal snoring sounds consist of only lower frequency components less than 500Hz, whereas oral snoring sounds consist of unique intensity peaks at 1kHz and lower frequency components less than 500Hz as well. According to the bibliographical point of view, such lower frequency components indicate the palatal snoring, and the intensity peak at around 1kHz indicates the tongue base snoring. Therefore, it is obvious that nasal snoring sounds can be regarded as simple palatal snoring whereas oral snoring sounds are a mixture of palatal and tongue base snoring sounds. So, we focused on the fundamental frequency and maximum of the amplitude spectrum in a specific band. In this paper, the Harmonic Product Spectrum (HPS) method is used for estimating the fundamental frequency and the k-Nearest Neighbor method is adopted for classifying oral/nasal snoring sounds. As a result, over 89% of snoring sounds are successfully classified under the four kinds of cross validation evaluations.