多模态睡眠呼吸暂停综合征检测算法研究及应用_毕业设计_论文.pdf

ABSTRACT

III

ABSTRACT

Sleep Apnea Syndrome (SAS) is a common sleep disorder that has a serious impact on

the quality of sleep and daily life of people. Polysomnography is a gold standard for the

diagnosis of sleep apnea syndrome, which is complex to implement, has a large number of

leads, and can easily affect the quality of sleep of patients during implementation. It has been

shown in the literature that characteristic parameters in the ECG signal can characterize SAS,

and many existing methods are based on single-channel ECG signals for SAS detection, but

(1) ECG signal and oxygen saturation signal preprocessing. In this thesis, firstly, the ECG

signal and oxygen saturation signal are segmented, each segment is 5 minutes; the median

filter is used to remove the baseline drift in the ECG signal; the 50Hz FIR trap is used to

remove the industrial frequency interference in the ECG signal; the Butterworth filter is

designed to remove the high frequency noise in the ECG signal; the moving average filter is

used to remove the abnormal values in the oxygen saturation signal.

(2) To investigate SAS characterization methods based on the fusion of multimodal

physiological signal features. In this thesis, the Pan-Tompkins algorithm was used to detect

the position of the R peak in the ECG signal, calculate the RR interval and heart rate variability

of the ECG signal, and extract the feature parameters that can characterize SAS on this basis;

广东工业大学硕士专业学位论文

IV

shedding layer to the network to prevent overfitting, which reduces the complexity of the

model and improves the detection of SAS on the data sets. The model uses the signal features

of fused ECG signal and blood oxygen saturation as input, and the accuracy obtained from the

detection of SAS on Apnea database is 86.0%, which is 3.4% more accurate than that of single-

channel ECG signal as input to the model; the accuracy obtained from the detection on

UCDDB database is 79.0%, which is 0.3% more accurate than that of single-channel ECG

signal as input to the model. The results show that it is more advantageous to use multimodal

feature fusion to detect SAS.

(4) Design and implementation of a sleep apnea syndrome monitoring system. The

system can continuously collect the ECG signal and blood oxygen saturation of patients

目录

V

目录

ABSTRACT ..........................................................................................................................III

目录 ......................................................................................................................................... V

CONTENTS ....................................................................................................................... VIII

第一章 绪论 .......................................................................................................................... 1

第三章 心电信号和血氧饱和度预处理 .............................................................................. 17

广东工业大学硕士专业学位论文

VI

第五章 基于改进的 LeNet-5 卷积神经网络的睡眠呼吸暂停检测 .................................. 35

总结与展望 ............................................................................................................................ 57

参考文献 ................................................................................................................................ 59

攻读学位期间取得与学位论文相关的成果 ........................................................................ 65

目录

致 谢 ...................................................................................................................................... 67