%0 Journal Article
%T 单通道脑电信号LSTM睡眠分期智能终端<br>Intelligent Terminal Based on Single Channel EEG Signal LSTM Sleep Staging
%A 陈振东
%A 杨飞帆
%J Computer Science and Application
%P 1156-1168
%@ 2161-881X
%D 2019
%I Hans Publishing
%R 10.12677/CSA.2019.96131
%X <div style=\"text-align:justify;\">
	<span style=\"font-size:14px;\">优质的睡眠，是人们在日常工作和生活保持精力充沛的前提和保证。提出一种依托APP和云端服务器的智能终端。设计了一种基于深度学习方法的全新睡眠阶段自动分析方法。该方法基于单通道脑电信号(EEG)，使用临床多导睡眠监测(PSG)得到的大规模数据集对网络进行建模，并将模型部署到云端服务器，利用其运算资源进行睡眠阶段自动分析。对比实验的结果显示，使用小规模数据集时，该方法整体准确度达到77.1%，kappa系数为0.68，性能优于同类型基于人工神经网络的睡眠分期方法。依靠精准的睡眠分期结果，智能终端能为用户提供实时脑电波形绘制、睡眠质量评估报告以及闹钟轻柔唤醒等功能。<br />
Adequate quality sleep is the premise and guarantee for people to keep energetic during daily life. In the paper, an intelligent terminal based on APP and cloud server is proposed. Also a new sleep stages automatic analysis method based on deep learning is discussed in the paper. Based on single channel EEG, the method uses the large dataset obtained by clinical polysomnography (PSG) to model the network and is deployed on the cloud server, which can be used for automatic sleep stages analysis. The results of the comparative experiment showed that the overall accuracy of this method reached 77.1% and the kappa coefficient was 0.68 when using small datasets, which is better than the similar sleep staging method based on other artificial neural networks. With the accurate sleep staging results, the intelligent terminal can provide users with real-time brain wave drawing, sleep quality assessment report, smart wake-up services and other functions.</span> 
</div>
%K 单通道脑电信号，睡眠分期，智能终端，LSTM，云端服务器<br>Single Channel EEG Signal
%K Sleep Staging
%K Intelligent Terminal
%K LSTM
%K Cloud Server
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=31040