TY - GEN
T1 - An Accurate and Low-power Sleep Apnea Detector in the Edge-IoT Platform
AU - Sayeed, Md Abu
AU - Carr, Patrick
N1 - Publisher Copyright:
© 2023 IEEE.
PY - 2023
Y1 - 2023
N2 - Sleep apnea (SA) is a serious sleep disorder that causes various diseases such as hypertension, type 2 diabetes, depression, and obesity. It may lead to serious health issues, including stroke if left untreated. The monitoring of SA through existing medical devices takes a lot of time and lacks the user's convenience. In this paper, an Internet of Things (IoT) enabled and automated SA detection technique has been presented that uses a deep-learning algorithm to analyze single-channel electrocardiogram (ECG) data and classify SA events from the specified epoch. SA detection is performed in two stages. In the first stage, a 1-sec epoch is extracted from the ECG signal, which is supplied for feature extraction. In the second stage, SA pattern is trained offline using a deep neural network (DNN), and real-time SA detection is performed during the testing phase. The widely available open-source dataset from PhysioNet is utilized for validation purposes. The proposed approach enhances the current state of the art and makes it a potential candidate for wearable medical devices.
AB - Sleep apnea (SA) is a serious sleep disorder that causes various diseases such as hypertension, type 2 diabetes, depression, and obesity. It may lead to serious health issues, including stroke if left untreated. The monitoring of SA through existing medical devices takes a lot of time and lacks the user's convenience. In this paper, an Internet of Things (IoT) enabled and automated SA detection technique has been presented that uses a deep-learning algorithm to analyze single-channel electrocardiogram (ECG) data and classify SA events from the specified epoch. SA detection is performed in two stages. In the first stage, a 1-sec epoch is extracted from the ECG signal, which is supplied for feature extraction. In the second stage, SA pattern is trained offline using a deep neural network (DNN), and real-time SA detection is performed during the testing phase. The widely available open-source dataset from PhysioNet is utilized for validation purposes. The proposed approach enhances the current state of the art and makes it a potential candidate for wearable medical devices.
KW - Electrocardiogram (ECG)
KW - Feature Reduction
KW - Internet of Things (IoT)
KW - Low-power Consumption
KW - Sleep Apnea
UR - http://www.scopus.com/inward/record.url?scp=85178508751&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85178508751&partnerID=8YFLogxK
U2 - 10.1109/AIBThings58340.2023.10292470
DO - 10.1109/AIBThings58340.2023.10292470
M3 - Conference contribution
AN - SCOPUS:85178508751
T3 - 2023 IEEE International Conference on Artificial Intelligence, Blockchain, and Internet of Things, AIBThings 2023 - Proceedings
BT - 2023 IEEE International Conference on Artificial Intelligence, Blockchain, and Internet of Things, AIBThings 2023 - Proceedings
A2 - Abdelgawad, Ahmed
A2 - Jamil, Akhtar
A2 - Hameed, Alaa Ali
T2 - 2023 IEEE International Conference on Artificial Intelligence, Blockchain, and Internet of Things, AIBThings 2023
Y2 - 16 September 2023 through 17 September 2023
ER -