深度学习方法在睡眠呼吸暂停检测中的研究进展

Research progress of deep learning methods in sleep apnea detection

睡眠呼吸暂停综合征是一种由上气道部分或完全阻塞引起的常见睡眠呼吸系统疾病,易诱发高血压、冠心病等心脑血管疾病,对人们的睡眠质量以及身心健康具有严重影响。近年来,深度学习方法在睡眠呼吸暂停检测中的应用研究受到了越来越多的关注。为推进基于深度学习的睡眠呼吸暂停检测技术的研究发展,论文对当前主流的基于深度学习的睡眠呼吸暂停检测方法进行了系统梳理和总结,介绍了常见的睡眠呼吸暂停检测公开数据集,给出了基于深度学习睡眠呼吸暂停检测方法演化发展过程,综述了近年来深度学习方法在睡眠呼吸暂停检测中的研究进展,分析了典型方法的思路和特点,给出了典型方法的实验比较,最后给出现阶段研究所存在的问题,并对未来研究及发展趋势进行了展望。

Sleep apnea syndrome is a common sleep respiratory disorder caused by partial or complete upper airway obstruction,which can lead to hypertension,coronary heart disease,and other cardiovascular diseases.It also seriously impacts sleep quality and physical and mental health.Sleep polysomnography monitoring is commonly used for determining and confirming sleep apnea,but manually analyzing sleep polysomnography is time-consuming,labor-intensive,and error-prone.In recent years,the research on the application of deep learning methods in sleep apnea detection has received increasing attention.To promote the research development of deep learning-based sleep apnea detection technology,this paper systematically composes and summarizes the current mainstream deep learning-based sleep apnea detection methods,introduces the common sleep apnea detection public datasets,describes the evolutionary development process of deep learning-based sleep apnea detection methods,reviews the deep learning methods in sleep apnea detection in recent years,analyzes the ideas and characteristics of typical methods,characterizes the experimental comparison of typical methods,appraises the problems existing at the present stage of research,and discusses future research and development trends.