Recognising Your Snore Sound: From Acoustic Parameter Analysis to Machine Learning based Methods

抄録

Snore sound, as a common symptom among adults (more than 30%), has been increasingly studied during the past three decades. Particularly, an in-depth analysis of snore sound can benefit a targeted surgical plan for both the subjects suffering from primary snoring and Obstructive Sleep Apnoea snoring. In this presentation, we will firstly introduce the history of using snore sound analysis for localising the snore site in the upper airway. We will compare the early works focused on acoustic parameter analysis and the machine learning (including the cutting-edge deep learning approaches) based studies by a comprehensive review on literature. Then, we indicate the findings and limitations in current research work. Finally, we conclude the studies and give our future perspectives.