As mobile and wearable devices become widespread, sensors and media input methods are used to track daily health records. Previously, the step counts, workouts, and sleep efficiencies are measured through accelerometer or GPS sensor in the device. It has been gradually expanded to other health information monitoring using various media input peripherals. For example, camera-based image recognition and voice input can be used as easy tracking methods. And optical sensors can continuously measure the various biometric information, such as heart rate, oxygen saturation, stress level, and even blood pressure. Using these various information, the mobile healthcare services can encourage users to engage in healthy activities or provide personalized recommendation. In this talk, I will introduce the related market trends and some challenges.

Obstructive Sleep Apnea (OSA) is a common, but severely under-diagnosed sleep disorder characterized by recurring periods of shallow or paused breathing during sleep. It is our long-term goal to allow people to perform the first step towards a sleep apnea detection at home by utilizing smartphones, low-cost consumer-grade sensors, and data mining techniques. In this work, we evaluate the signal quality of four respiratory effort sensors (BITalino, FLOW, RespiBAN, and Shimmer), using a RIP sensor from NOX Medical as the gold standard. We design a sixteen-minute signal capture procedure to simulate epochs of disrupted breathing, and capture data from twelve (BITalino and Shimmer) and eleven (RespiBAN and FLOW) subjects during wakefulness. Our signal quality evaluation approach is based on the breath detection accuracy metrics sensitivity and positive predictive value (PPV), along with the breath amplitude accuracy metric weighted absolute percentage error (WAPE). These metrics are closely related to how apneic and hypopneic episodes are scored by medical personnel, making it straightforward to reason about their interpretation. Our results show that false breaths are the primary concern affecting the breath detection accuracy of BITalino, Shimmer, and RespiBAN. Respectively, the sensitivity of BITalino, Shimmer, RespiBAN, and FLOW is 99.61%, 98.53%, 98.41%, and 98.91%. Their PPV is 96.28%, 96.58%, 90.81%, and 98.81%. Finally, their WAPE is 13.82%, 16.89%, 13.60%, and 8.75%. The supine (back) position is consistently showing the overall best signal quality compared to the side position.

Cardiovascular disease (CVD) is the leading cause of premature death and disability in Europe and worldwide. Effective Cardiac Rehabilitation (CR) can significantly improve mortality and morbidity rates, leading to longer independent living and a reduced use of health care resources. However, adherence to such an exercise programme is generally low for a variety of reasons such as lack of time and how enjoyable the CR programme is. In this work, we proposed a method for automatic enjoyment estimation during an exercise which could be used by a clinician to identify when a patient is not enjoying the exercise and therefore at risk of early dropout. In order to evaluate the proposed method, a database was captured where participants perform various of CR exercises. Three set of facial features were extracted and were evaluated using seven different classifiers. The proposed method achieved 49% average accuracy in predicting five different enjoyment level on the newly collected database.

This study investigated the differences in the psychophysiological reaction to anxious situations in individuals with higher (greater than or equal to 40) trait anxiety scores in comparison to those with lower (less than 40) trait anxiety scores. This information may be useful for convenient anxiety treatment options and health trackers toward effectively recording and interpreting physiological data from individuals who are generally more anxious. Ten adults completed an IRB (Institutional Review Board) approved study in which all participants completed the Trait Anxiety Scale (Form Y-2) of the State Trait Anxiety Inventory, and subsequently underwent four phases of baseline, introduction, virtual reality simulation, and recovery during which EEG (electroencephalogram), heart rate, and skin conductance data was collected. Participants also recorded their self-interpreted anxiety on a scale of 1-10 after each phase of the experiment. The brief introduction phase and virtual reality simulation were designed to elicit mild anxiety. Results show no statistically significance difference in average percent difference in skin conductance or heart rate changes between baseline to introduction, baseline to virtual reality or baseline to recovery between individuals with high (greater than or equal to 40) trait anxiety scores and average or low (less than 40) scores. These findings imply important information that trait anxiety does not necessarily correlate to more severe physiological reactions to anxious situations and confirms that manifestations of anxiety may vary greatly between individuals. Most importantly, evaluative measures for the effectiveness of potential health tracking applications or anxiety treatments would be most effective if perceived anxiety intensities were given more value than solely physiological data.

ADHD has an estimated worldwide prevalence of 2-3% and is one of the most frequent neurodevelopmental disorders. The disease starts in childhood and persists in up to 50% into adulthood. Many problems in an ADHD-patient's life arise from the lack of self-management abilities. Technologies may be able to support adolescents and adults with ADHD by strengthening self-awareness of their symptoms, monitoring treatment and helping to develop responsibility, emotional self-regulation and self-management. In our work, we develop a modular on-body system which supports self-management of psychological disorders for young adults unobtrusively in everyday life. We present insights into the future concept of an assessment system based on contextual experience sampling and biomarkers, to predict agitated situations and enable patients to reflect on their own behaviors.

Activity recognition is important to health care in smart homes. It provides information about the activities of the residents. Many health care services are based on it. To collect data about the activities, sensor networks consist of binary sensors are widely used. Activity recognition is performed based on their readings. Most of existing activity recognition methods are based on supervised classification algorithms. One drawback of these methods is that the classification model learned in one smart home environment usually cannot be used in another. For a new smart home environment, sufficient sensor readings have to be collected and labeled to learn the needed classification model. This process is time consuming and expensive. In this paper, we propose a method for smart home activity recognition with binary sensors. Our method utilizes the characteristics of binary sensors, the semantic information of the sensors and the activities, and the time information. The classification model learned with the data in one smart home environment can be used in the activity recognition in another, which has different sensor networks and label spaces. Experiments on real world datasets show the effectiveness of our method.

A food journal is essential for improving health and well-being. However, journaling every meal is extremely difficult because it depends on user initiative and intervention. Current approaches to food journaling are both potentially inaccurate and tedious, causing people to abandon their journals very soon after they start. In this paper, we propose a proactive and reactive mechanism that can significantly reduce user initiative while still remaining highly accurate. We first suggest a novel eating moment recognition technique using heart rate and activity patterns to trigger a food journaling process in a proactive manner. We then begin the food journaling process via voice command which utilized natural language processing when logging meals, which increases the ease of reactive self-reporting. Lastly, we enhance the food journal by automatically assessing ecological moments of eating activity through our personal chronicle system. We verified the method from a feasibility study conducted with three people for three months in their day-to-day lives. Our approach is designed to be unobtrusive and practical by leveraging multi-modal sensor data through the most common device combination of a smartphone and wearable device.

Older adults have been encountering difficulties in using modern technological devices to control home appliances as they are lacking in technology literacy and mobility. This led to the usage of remote controllers or requiring assistance from family members, which is not beneficial for older adults since there is less independence. To alleviate this problem, this project aims to develop a prototype system named "Genie" which caters for older adults ranging from 65 to 80 years old, allowing for easy control of smart home appliances through combination of speech and gesture interactions. An experiment was carried out with a total of 20 older adults on the prototype system where the initial results demonstrate a significant increase in usability. Based on the evaluation, such interaction methods show promise to be effective in replacing manual operations of home appliances through the use of simple speech or gesture commands.

Stair climbing is a physical activity that can easily be performed in daily life and has a positive influence on, amongst others, cardiovascular health and the prevention of frailty. Health interventions have shown to be effective in motivating stair climbing in public spaces, but have so far mostly relied on analogue means such as banners or posters. We investigate the role of technology to promote stair climbing in public spaces. We designed the ActiStairs system, a simple and practical system to be used in real-life circumstances. To understand acceptance we observed the users' reaction on the system in a public shopping mall. Based on our findings we suggest implications for future technology for the promotion of stair climbing in public spaces.

Therapists and trainers often utilize training and exercise programs as a critical step within the rehabilitation process. Due to the inherent danger of injury while performing these exercises, they must often be performed under the supervision of skilled trainers to observe and provide feedback on body compliance during the motion. To address this safety risk, we've developed the Smart-Gym: an intelligent modification to the Total Gym Pro that monitors a user's body during exercise and provides feedback to anticipate non-compliance. The system was built to provide multimodal feedback for various body adjustments in the same way that a trainer would. Results from an initial case study involving an individual undergoing exercise rehabilitation for hemiparetic cerebral palsy are discussed.