Anxiety disorders are the most common class of psychiatric problems affecting both children and adults. However, tools to effectively monitor and manage anxiety are lacking, and comparatively limited research has been applied to addressing the unique challenges around anxiety. Leveraging passive and unobtrusive data collection from smartphones could be a viable alternative to classical methods, allowing for real-time mental health surveillance and disease management. This paper presents eWellness, an experimental mobile application designed to track a full-suite of sensor and user-log data off an individual’s device in a continuous and passive manner. We report on an initial pilot study tracking ten people over the course of a month that showed a nearly 76% success rate at predicting daily anxiety and depression levels based solely on the passively monitored features.
The majority of the population around the globe spend a considerable portion of their days seated. This fact can be associated with several factors, such as the circumstances of most of the current jobs and the prevalence of the use of computer systems. One could argue that this knowledge indicates that the impacts of maintaining proper posture while sitting can be observed more than before. Therefore, it is critical to be able to observe, correct, and control our sitting posture throughout the day. Monitoring and correcting our short-term and longterm sitting habits over time can lead to significant improvement in our physical well-being. In this work, we propose WatChair, an AI-powered remote subject monitoring system that assists in short-term sitting posture recognition, activity-level tracking, long-term monitoring, and providing corrective suggestions. Our platform consists of a small wearable component, an application, and a cloud-based back-end. Our framework has been evaluated in practice, and the results of empirical validation and the user-friendliness questionnaire points to a simple, accurate, and user-friendly system for remote sitting posture monitoring. This framework also presents an adaptable solution for general dynamic posture recognition and tracking using wearable systems based on motion sensors.
Developing and maintaining monitoring panels is undoubtedly the main task in the remote patient monitoring (RPM) systems. Due to the significant variations in desired functionalities, data sources, and objectives, designing an efficient dashboard that responds to the various needs in an RPM project is generally a cumbersome task to carry out. In this work, we present ViSierra, a framework for designing data monitoring dashboards in RPM projects. The abstractions and different components of this open-source project are explained, and examples are provided to support our claim concerning the effectiveness of this framework in preparing fast, efficient, and accurate monitoring platforms with minimal coding. These platforms will cover all the necessary aspects in a traditional RPM project and combine them with novel functionalities such as machine learning solutions and provide better data analysis instruments for the experts to track the information.
Completing a set of therapeutic exercises correctly and regularly is essential for early and safe recovery from knee reconstructive surgery. We propose EXTRA, a platform for monitoring and improving the quality of therapeutic exercises that the patients need to perform during the recovery. The data acquisition framework obtains observations of the knee movement using an embedded flex sensor in a hinged knee brace. A connected Android application transmits the data to the database real-time. We evaluated the accuracy of the knee joint angle measurement and the usability in a study with young healthy adults. The mean absolute error of the knee angle measurements was 13.31 degrees. The high score of the usability questionnaire indicates that EXTRA provides a user-friendly interface to motivate therapeutic exercises. EXTRA maintains high standards in accurately monitoring and interpreting the exercises. Our proposed platform has the potential to provide effective guidelines and improve the rehabilitation process for orthopedic professionals and patients.
In this paper, we study the problem of children activity recognition using smartwatch devices. We introduce the need for a robust children activity model and challenges involved. To address the problem, we employ two deep neural network models, specifically, Bi-Directional LSTM model and a fully connected deep network and compare the results to commonly used models in the area. We demonstrate that our proposed deep models can significantly improve results compared to baseline models. We further show benefits of activity intensity level detection in health monitoring and verify high performance of our proposed models in this task.