Large-Scale Display Interaction Techniques to Support Face-to-Face Collaboration
Thesis DisciplineComputer Science
Degree GrantorUniversity of Canterbury
Degree NameMaster of Science
This research details the development of a large-scale, computer vision-based touch screen capable of supporting a large number of simultaneous hand interactions. The system features a novel lightweight multi-point tracking algorithm to improve real-time responsiveness. This system was trialled for six months in an exhibition installation at World Expo 2005 in Aichi, Japan, providing a robust, fault-tolerant interface. A pilot study was then conducted to directly compare the system against other, more established input methods (a single-touch case, a two-mouse case and a physical prototype) to determine the effectiveness and affordances of the multi-touch technology for arranging information on a large-scale wall space in a paired collaborative task. To assist in this study, a separate visualisation and interaction classification tool was developed, allowing the replay of XML log data in real time to assist in the video analysis required for observation and hypothesis testing.