Paper Reading #2: Pinch-the-sky dome: freehand multi-point interactions with immersive omni-directional data

Comments
Luke Roberts
Cindy Skach

Reference Information

Title: Pinch-the-sky dome: freehand multi-point interactions with immersive omni-directional data
Authors: Hrvoje Benko, Microsoft Research, Redmon, A, USA
              Andrew D. Wilson, Microsoft Research, Redmon, A, USA
Presentation Venue: CHI 2010: 28th ACM Conference on Human Factors in Computing Systems;
Date: April 10-15, 2010;
Location: Atlanta, GA, USA

Summary

Pinch-the-Sky is an immersive installation that allows users to interact simultaneously with omni-directional data projected inside the walls of a cardboard geodesic dome. In the middle of the room there is an omni-directional projector able to project 360 degrees. With the projector there is a camera that can sense hand gestures over the projector.
The types of activities provided to the user are:

1. The exploration of the astronomical data provided by World Wide Telescope
2. Social networking 3D graph visualizations
3. Immersive panoramic images
4. 360 degree video conferencing 

With a specific interaction, the user can move the view with their hands, pinch-to-zoom similar to Microsoft Surface. There is also voice recognition implementated.

In implementing this solution, two main problems emerge:

1.  The dome is supposed to be usable by more than one user, and in that lies the difficulty of making distictions between voluntary and involuntary gestures.
2. The second problem lies with the camera making the distinction between a user actively engaging with the display, or merely standing by as a spectator.

Discussion

This paper was quite interesting. It does a good job of introducing the subject, talking about the hardware being used, and the method of interaction. It also discusses some of the challenges faced. In it we learn about the pinch-to-zoom technique borrowed from the Microsoft Surface. If I were working on this project, I would try to implement a simple hand signal that the user could make to let the system know he wants to interact with the system. Perhaps that could resolve the hard task of distinguishing between a natural involutary and volutary gesture.