論文アブストラクト： We present a data logging concept, tool, and analyses to facilitate studies of everyday mobile touch keyboard use and free typing behaviour: 1) We propose a filtering concept to log typing without recording readable text and assess reactions to filters with a survey (N=349). 2) We release an Android keyboard app and backend that implement this concept. 3) Based on a three-week field study (N=30), we present the first analyses of keyboard use and typing biometrics on such free text typing data in the wild, including speed, postures, apps, auto correction, and word suggestions. We conclude that research on mobile keyboards benefits from observing free typing beyond the lab and discuss ideas for further studies.
論文アブストラクト： We introduce the concept of Veritaps: a communication layer to help users identify truths and lies in mobile input. Existing lie detection research typically uses features not suitable for the breadth of mobile interaction. We explore the feasibility of detecting lies across all mobile touch interaction using sensor data from commodity smartphones. We report on three studies in which we collect discrete, truth-labelled mobile input using swipes and taps. The studies demonstrate the potential of using mobile interaction as a truth estimator by employing features such as touch pressure and the inter-tap details of number entry, for example. In our final study, we report an F1-score of .98 for classifying truths and .57 for lies. Finally we sketch three potential future scenarios of using lie detection in mobile applications; as a security measure during online log-in, a trust layer during online sale negotiations, and a tool for exploring self-deception.
論文アブストラクト： Learning-based gaze estimation has significant potential to enable attentive user interfaces and gaze-based interaction on the billions of camera-equipped handheld devices and ambient displays. While training accurate person- and device-independent gaze estimators remains challenging, person-specific training is feasible but requires tedious data collection for each target device. To address these limitations, we present the first method to train person-specific gaze estimators across multiple devices. At the core of our method is a single convolutional neural network with shared feature extraction layers and device-specific branches that we train from face images and corresponding on-screen gaze locations. Detailed evaluations on a new dataset of interactions with five common devices (mobile phone, tablet, laptop, desktop computer, smart TV) and three common applications (mobile game, text editing, media center) demonstrate the significant potential of cross-device training. We further explore training with gaze locations derived from natural interactions, such as mouse or touch input.
論文アブストラクト： Smartwatches enable rapid access to information anytime and anywhere. However, current smartwatch content navigation techniques, for panning and zooming, were directly adopted from those used on smartphones. These techniques are cumbersome when performed on small smartwatch screens and have not been evaluated for their support in mobility and encumbrance contexts (when the user's hands are busy). We studied the effect of mobility and encumbrance on common content navigation techniques and found a significant decrease in performance as the pace of mobility increases or when the user was encumbered with busy hands. Based on these initial findings, we proposed a design space which would improve efficiency when navigation techniques, such as panning and zooming, are employed in mobility contexts. Our results reveal that our design space can effectively be used to create novel interaction techniques that improve smartwatch content navigation in mobility and encumbrance contexts.