Abstract:
Interactive tabletop systems are becoming popular platforms for group activities. However, current common tabletops do not provide capabilities to differentiate interactions among simultaneous users, i.e. to associate a touch point with its proper owner. My thesis proposes and explores the use of an important biometric property of users as the basis for touch discrimination on multi-user tabletops: Finger Orientation (FO). In this thesis, I first collect the FO ranges of users standing in different positions around a tabletop. Second, I implement a system that uses FO to determine where the users are standing, and based on that extrapolate the owner of the touch. Next, I evaluate the system with two separate experiments, present the results, and discuss all findings. Furthermore, I explore some enhancements with a simple quantitative study. My results indicate that finger orientation is a good natural biometric trait enhances multi-user recognition on tabletops.
