|Course||Pervasive Computing Methodenseminar - Taxonomy of Bypassers|
|Audience||Students of Computer Science.|
|Goals||An in depth investigation of selected topics in computer science, in particular topics related to Embedded and Pervasive Systems.
Improving skill and competence in literature research, independent work on a topic, and presentation capabilities.
|Contents||Interactive public displays use cameras, depth sensors, microphones and several other sensors to inspect its audience. Gesture recognition or even simple properties like the distance of an interacting person to the display can trigger specific responses. But even seemingly simple properties often require a complex classification process, involving several sensors. Taxonomies for activities, mood or attention have to be devised and methods found to classify bypassers accordingly.
Classification of bypassers of a public display according to their attention level. 
In this seminar you will have to investigate what public displays can perceive, which sensors are available and what kind of behavioural patterns can be detected in bypassers. You will have to create a taxonomy of bypassers, then design and execute a user study to validate your assumptions. Experiments will be performed with the Microsoft Kinect or other sensors mounted in the "media marterl" in the PC Lab.
Image of the Kinect depth sensor, with highlighted bypassers. 
 B. Gollan, B. Wally, A. Ferscha, "Automatic Attention Estimation in an Interactive System based on Behaviour Analysis", Proceedings of the 15th Portuguese Conference on Artificial Intelligence (EPIA2011), 2011.
|Literature||B. Gollan, B. Wally, A. Ferscha, Automatic Attention Estimation in an Interactive System based on Behaviour Analysis (see "Begleitmaterial" for pdf download)|