PM10 and PM2.5 real-time prediction models using an interpolated convolutional neural network.

In this paper, we propose a real-time prediction model that can respond to particulate matters (PM) in the air, which are an indication of poor air quality. The model applies interpolation to air quality and weather data and then uses a Convolutional Neural Network (CNN) to predict PM concentrations. The interpolation transforms the irregular spatial data into an equally spaced grid, which the model requires. This combination creates the interpolated CNN (ICNN) model that we use to predict PM10 and PM2.5 concentrations. The PM10 and PM2.5 evaluation results show an effective prediction performance with an R-squared higher than 0.97 and a root mean square error (RMSE) of approximately 16% of the standard deviation. Furthermore, both PM10 and PM2.5 prediction models forecast high concentrations with high reliability, with a probability of detection higher than 0.90 and a critical success index exceeding 0.85. The proposed ICNN prediction model achieves a high prediction performance using spatio-temporal information and presents a new direction in the prediction field.

Predicting Infectious Disease Using Deep Learning and Big Data.

Infectious disease occurs when a person is infected by a pathogen from another person or an animal. It is a problem that causes harm at both individual and macro scales. The Korea Center for Disease Control (KCDC) operates a surveillance system to minimize infectious disease contagions. However, in this system, it is difficult to immediately act against infectious disease because of missing and delayed reports. Moreover, infectious disease trends are not known, which means prediction is not easy. This study predicts infectious diseases by optimizing the parameters of deep learning algorithms while considering big data including social media data. The performance of the deep neural network (DNN) and long-short term memory (LSTM) learning models were compared with the autoregressive integrated moving average (ARIMA) when predicting three infectious diseases one week into the future. The results show that the DNN and LSTM models perform better than ARIMA. When predicting chickenpox, the top-10 DNN and LSTM models improved average performance by 24% and 19%, respectively. The DNN model performed stably and the LSTM model was more accurate when infectious disease was spreading. We believe that this study's models can help eliminate reporting delays in existing surveillance systems and, therefore, minimize costs to society.

Sinabro: A Smartphone-Integrated Opportunistic Electrocardiogram Monitoring System.

In our preliminary study, we proposed a smartphone-integrated, unobtrusive electrocardiogram (ECG) monitoring system, Sinabro, which monitors a user's ECG opportunistically during daily smartphone use without explicit user intervention. The proposed system also monitors ECG-derived features, such as heart rate (HR) and heart rate variability (HRV), to support the pervasive healthcare apps for smartphones based on the user's high-level contexts, such as stress and affective state levels. In this study, we have extended the Sinabro system by: (1) upgrading the sensor device; (2) improving the feature extraction process; and (3) evaluating extensions of the system. We evaluated these extensions with a good set of algorithm parameters that were suggested based on empirical analyses. The results showed that the system could capture ECG reliably and extract highly accurate ECG-derived features with a reasonable rate of data drop during the user's daily smartphone use.

Unconstrained detection of freezing of Gait in Parkinson's disease patients using smartphone.

Freezing of gait (FOG) is a common motor impairment to suffer an inability to walk, experienced by Parkinson's disease (PD) patients. FOG interferes with daily activities and increases fall risk, which can cause severe health problems. We propose a novel smartphone-based system to detect FOG symptoms in an unconstrained way. The feasibility of single device to sense gait characteristic was tested on the various body positions such as ankle, trouser pocket, waist and chest pocket. Using measured data from accelerometer and gyroscope in the smartphone, machine learning algorithm was applied to classify freezing episodes from normal walking. The performance of AdaBoost.M1 classifier showed the best sensitivity of 86% at the waist, 84% and 81% in the trouser pocket and at the ankle respectively, which is comparable to the results of previous studies.

ROI analysis for remote photoplethysmography on facial video.

As wide spreading of camera-equipped devices to the daily living environment, there are enormous opportunities to utilize the camera-based remote photoplethysmography (PPG) for daily physiological monitoring. In the camera-based remote PPG (rPPG) monitoring, the region of interest (ROI) is related to the signal quality and the computational load for the signal extraction processing. Designating the best ROI on the body while minimizing its size is essential for computationally efficient rPPG extraction. In this study, we densely analyzed the face region to find the computationally efficient ROI for facial rPPG extraction. We divided the face into seven regions and evaluated the quality of the signal of each region using the area ratio of high-SNR and high-correlation, and mean and standard deviation (SD) of SNR and correlation coefficient. The results show that a forehead and both cheeks especially have a potential to be a good candidates for computationally efficient ROI. On the other hand, the signal quality from a mouth and a chin was relatively low. A nasion and a nose have a limitation to be efficient ROI.

CardioGuard: a brassiere-based reliable ECG monitoring sensor system for supporting daily smartphone healthcare applications.

We propose CardioGuard, a brassiere-based reliable electrocardiogram (ECG) monitoring sensor system, for supporting daily smartphone healthcare applications. It is designed to satisfy two key requirements for user-unobtrusive daily ECG monitoring: reliability of ECG sensing and usability of the sensor. The system is validated through extensive evaluations. The evaluation results showed that the CardioGuard sensor reliably measure the ECG during 12 representative daily activities including diverse movement levels; 89.53% of QRS peaks were detected on average. The questionnaire-based user study with 15 participants showed that the CardioGuard sensor was comfortable and unobtrusive. Additionally, the signal-to-noise ratio test and the washing durability test were conducted to show the high-quality sensing of the proposed sensor and its physical durability in practical use, respectively.

Highly wearable galvanic skin response sensor using flexible and conductive polymer foam.

Owing to advancements in daily physiological monitoring technology, diverse healthcare applications have emerged recently. The monitoring of skin conductance responses has extensive feasibility to support healthcare applications such as detecting emotion changes. In this study, we proposed a highly wearable and reliable galvanic skin response (GSR) sensor that measures the signals from the back of the user. To enhance its wearability and usability, we employed flexible conductive foam as the sensing material and designed it to be easily attachable to (and detachable from) a wide variety of clothes. We evaluated the sensing reliability of the proposed sensor by comparing its signal with a reference GSR. The average correlation between the two signals was 0.768; this is sufficiently high to validate the feasibility of the proposed sensor for reliable GSR sensing on the back.

Unobtrusive monitoring of ECG-derived features during daily smartphone use.

Heart rate variability (HRV) is known to be one of the representative ECG-derived features that are useful for diverse pervasive healthcare applications. The advancement in daily physiological monitoring technology is enabling monitoring of HRV in people's everyday lives. In this study, we evaluate the feasibility of measuring ECG-derived features such as HRV, only using the smartphone-integrated ECG sensors system named Sinabro. We conducted the evaluation with 13 subjects in five predetermined smartphone use cases. The result shows the potential that the smartphone-based sensing system can support daily monitoring of ECG-derived features; The average errors of HRV over all participants ranged from 1.65% to 5.83% (SD: 2.54~10.87) for five use cases. Also, all of individual HRV parameters showed less than 5% of average errors for the three reliable cases.

Validation of heart rate extraction using video imaging on a built-in camera system of a smartphone.

As a smartphone is becoming very popular and its performance is being improved fast, a smartphone shows its potential as a low-cost physiological measurement solution which is accurate and can be used beyond the clinical environment. Because cardiac pulse leads the subtle color change of a skin, a pulsatile signal which can be described as photoplethysmographic (PPG) signal can be measured through recording facial video using a digital camera. In this paper, we explore the potential that the reliable heart rate can be measured remotely by the facial video recorded using smartphone camera. First, using the front facing-camera of a smartphone, facial video was recorded. We detected facial region on the image of each frame using face detection, and yielded the raw trace signal from the green channel of the image. To extract more accurate cardiac pulse signal, we applied independent component analysis (ICA) to the raw trace signal. The heart rate was extracted using frequency analysis of the raw trace signal and the analyzed signal from ICA. The accuracy of the estimated heart rate was evaluated by comparing with the heart rate from reference electrocardiogram (ECG) signal. Finally, we developed FaceBEAT, an iPhone application for remote heart rate measurement, based on this study.

Validation of heart rate extraction through an iPhone accelerometer.

Ubiquitous medical technology may provide advanced utility for evaluating the status of the patient beyond the clinical environment. The iPhone provides the capacity to measure the heart rate, as the iPhone consists of a 3-axis accelerometer that is sufficiently sensitive to perceive tiny body movements caused by heart pumping. In this preliminary study, an iPhone was tested and evaluated as the reliable heart rate extractor to use for medical purpose by comparing with reference electrocardiogram. By comparing the extracted heart rate from acquired acceleration data with the extracted one from ECG reference signal, iPhone functioning as the reliable heart rate extractor has demonstrated sufficient accuracy and consistency.