http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
A Binary Sleep Classification of BCG Signal using Random Search-Optimized Random Forest
Unang Sunarya,Hyunsoo Yu,Sayup Kim,Cheolsoo Park 대한전자공학회 2021 대한전자공학회 학술대회 Vol.2021 No.6
Sleep has an important role in human daily activities. Lack of sleep might cause fatigue, loss of focus, and furthermore lead to various sleep-related diseases. Research on human sleep has been conducted in many ways starting from high-cost and complex process (polysomnography) to low-cost and simpler process such as actigraphy. Several different methods were used during the experiment such as (EEG), electrocardiography (ECG), and ballistocardiography (BCG) to automate the process for sleep analysis. A machine learning algorithm called random forest has been used to classify awake and sleep states. Additionally, the random forest algorithm was optimized with random search to improve the classification performances. The result showed the accuracy of the optimized random forest was higher (81.35 %) than the result without using the optimization method (79.82 %).
Optimal Number of Cardiac Cycles for Continuous Blood Pressure Estimation
Unang Sunarya,Cheolsoo Park 대한전자공학회 2022 IEIE Transactions on Smart Processing & Computing Vol.11 No.6
Continuous blood pressure monitoring is essential for patients with hypertension. Most studies have suggested cuffless blood pressure monitoring techniques using a single cardiac cycle based on the pulse transit time. This paper investigates feature extraction from multiple cardiac cycles to estimate blood pressure. This implementation uses electrocardiogram, photoplethysmogram, and ballistrocardiagram signals. Random forest was applied to estimate blood pressure using the leave-one-subject-out cross-validation technique. The results show that the model could achieve better performance when using three cardiac cycles with an average MAE of 3.364 ± 3.059 mmHg for the diastolic blood pressure and 4.201 ±2.355 mmHg for the systolic blood pressure.
Continuous Blood Pressure Estimation using 1D Convolutional Neural Network and Attention Mechanism
Youjung Seo,Jungwhan Lee,Unang Sunarya,Kwangkee Lee,Cheolsoo Park 대한전자공학회 2022 IEIE Transactions on Smart Processing & Computing Vol.11 No.3
Patients with hypertensive blood pressure (BP) needs a round-the-clock BP monitoring and must take precautions to prevent emergencies such as stroke or heart failure. This paper suggests a deep neural network (DNNs–based BP estimation approach using electrocardiogram (ECG), photoplethysmogram (PPG), and ballistocardiogram (BCG) signals. The proposed approach consists of a one-dimensional convolutional neural network (1D CNN) followed by the attention mechanism known as Luong attention. Estimations under the proposed model yield mean absolute error (MAE) of 3.299±2.419 for systolic and 2.69±1.821 for diastolic BP. The algorithm can effectively predict BP without a recurrent neural network (RNNs), which is a typical DNNs model for processing sequential data. Additionally, the proposed approach is preferable owing to its ability to explain the model.
Blood pressure estimation and its recalibration assessment using wrist cuff blood pressure monitor
Youjung Seo,Saehim Kwon,Unang Sunarya,Sungmin Park,Kwangsuk Park,Dawoon Jung,Youngho Cho,Cheolsoo Park 대한의용생체공학회 2023 Biomedical Engineering Letters (BMEL) Vol.13 No.2
The rapid evolution of wearable technology in healthcare sectors has created the opportunity for people to measure theirblood pressure (BP) using a smartwatch at any time during their daily activities. Several commercially-available wearabledevices have recently been equipped with a BP monitoring feature. However, concerns about recalibration remain. Pulsetransit time (PTT)-based estimation is required for initial calibration, followed by periodic recalibration. Recalibration usingarm-cuff BP monitors is not practical during everyday activities. In this study, we investigated recalibration using PTT-basedBP monitoring aided by a deep neural network (DNN) and validated the performance achieved with more practical wristcuffBP monitors. The PTT-based prediction produced a mean absolute error (MAE) of 4.746 ± 1.529 mmHg for systolicblood pressure (SBP) and 3.448 ± 0.608 mmHg for diastolic blood pressure (DBP) when tested with an arm-cuff monitoremploying recalibration. Recalibration clearly improved the performance of both DNN and conventional linear regressionapproaches. We established that the periodic recalibration performed by a wrist-worn BP monitor could be as accurate asthat obtained with an arm-worn monitor, confirming the suitability of wrist-worn devices for everyday use. This is the firststudy to establish the potential of wrist-cuff BP monitors as a means to calibrate BP monitoring devices that can reliablysubstitute for arm-cuff BP monitors. With the use of wrist-cuff BP monitoring devices, continuous BP estimation, as wellas frequent calibrations to ensure accurate BP monitoring, are now feasible.