Sleep is a crucial factor in our daily life, and it is defined as behavior that maintains an unconscious state. In general, adult people require around 7 to 8 hours of sleep to rest their mind and body properly, thereby relieving stress and helping to prevent various diseases. Over the past few deca...
Sleep is a crucial factor in our daily life, and it is defined as behavior that maintains an unconscious state. In general, adult people require around 7 to 8 hours of sleep to rest their mind and body properly, thereby relieving stress and helping to prevent various diseases. Over the past few decades, research has shown that sleep time is gradually declining. The decrease in sleeping time adversely affects people's quality of life by inducing fatigue, memory and concentration reduction, hypertension, diabetes, and obesity. Considering the effects of sleep deprived, research interests in sleep have increased recently, in like manner understanding and analysis of the sleep stage are becoming more important. Polysomnography is a method for measuring sleep stages. Polysomnography is performed in a hospital; it uses more than 22 sensors to read and record the changes in a body overnight while sleeping. To perform the polysomnography, the tester must attach the sensors that create an unfamiliar environment; hence, it is impossible to get an accurate result because of stress due to the unfamiliar environment. In addition, there is a possibility that the reading result differs from one hospital to another, depending on the experience of the reader. As a matter of fact, it takes a lot of time and cost to read the test results directly by comparing the recorded signals with each other. Therefore, in this study, we have studied a sleep stage classification algorithm based on machine learning that can replace polysomnography. We compare the performance of the EEG signal, EOG signal, and EMG signal by using CNN algorithm and compare CNN and DNN model to compensate low accuracy of light sleep stage. A hybrid model was designed.
Sleep is a crucial factor in our daily life, and it is defined as behavior that maintains an unconscious state. In general, adult people require around 7 to 8 hours of sleep to rest their mind and body properly, thereby relieving stress and helping to prevent various diseases. Over the past few decades, research has shown that sleep time is gradually declining. The decrease in sleeping time adversely affects people's quality of life by inducing fatigue, memory and concentration reduction, hypertension, diabetes, and obesity. Considering the effects of sleep deprived, research interests in sleep have increased recently, in like manner understanding and analysis of the sleep stage are becoming more important. Polysomnography is a method for measuring sleep stages. Polysomnography is performed in a hospital; it uses more than 22 sensors to read and record the changes in a body overnight while sleeping. To perform the polysomnography, the tester must attach the sensors that create an unfamiliar environment; hence, it is impossible to get an accurate result because of stress due to the unfamiliar environment. In addition, there is a possibility that the reading result differs from one hospital to another, depending on the experience of the reader. As a matter of fact, it takes a lot of time and cost to read the test results directly by comparing the recorded signals with each other. Therefore, in this study, we have studied a sleep stage classification algorithm based on machine learning that can replace polysomnography. We compare the performance of the EEG signal, EOG signal, and EMG signal by using CNN algorithm and compare CNN and DNN model to compensate low accuracy of light sleep stage. A hybrid model was designed.
주제어
#Sleep Stage Classification CNN algorithm DNN algorithm Polysomnography EEG Signal EOG Signal EMG Signal
※ AI-Helper는 부적절한 답변을 할 수 있습니다.