Tag-less feature based tangible detection and tracking using custom gestures

Abstract

Digital tabletops have gained prominence in research and industry, as they support new types of interaction, such as tangible user interfaces. The task of tangible detection and tracking is generally accomplished through using tags to mark the objects, or by forwarding information on the objects to developers, who do the identification and tracking. This thesis presents a methodology and toolkit through which tag-less object detection can be achieved, by using a feature-based computer vision algorithm. This allows developers to access their tangibles through the names that they provide and protects them from the complexities of computer vision algorithms and the tracking of tangibles. It also allows for the integration of multi-touch and tangible interaction, as both modalities are crucial to communicating with digital tabletops. Custom gestures can be defined which encompass touches and tangibles. A user study was conducted which demonstrated that participants found the toolkit to be useful and easy to use.