FlowPut

Abstract

Tangible interaction has shown to be beneficial in a wide variety of scenarios since it provides more direct manipulation and haptic feedback. Further, inherently three-dimensional information is represented more naturally by a 3D object than by a flat picture on a screen. Yet, today's tangibles have often pre-defined form factors and limited input and output facilities. To overcome this issue, the combination of projection and depth cameras is used as a fast and flexible way of non-intrusively adding input and output to tangibles. However, tangibles are often quite small and hence the space for output and interaction on their surface is limited. Therefore, we propose FlowPut: an environment-aware framework that utilizes the space available on and around a tangible object for projected visual output. By means of an optimization-based layout approach, FlowPut considers the environment of the objects to avoid interference between projection and real-world objects. Moreover, we contribute an occlusion resilient object recognition and tracking for tangible objects based on their 3D model and a point-cloud based multi-touch detection, that allows sensing touches also on the side of a tangible. Flowput is validated through a series of technical experiments, a user study, and two example applications.