Reference Information
Title: A screen-space formulation for 2D and 3D direct manipulation
Title: A screen-space formulation for 2D and 3D direct manipulation
Authors: Jason L. Reisman, Perceptive Pixel, Inc., New York, NY, USA
Philip L. Davidson, Perceptive Pixel, Inc., New York, NY, USA
Jefferson Y. Han, Perceptive Pixel, Inc., New York, NY, USA
Presentation Venue: UIST 2009: 22nd annual ACM symposium on User interface software and technology;
Date: 2009;
Location: New York, NY, USA
Summary
Date: 2009;
Location: New York, NY, USA
Summary
This paper presents a new approach to touch-screen interaction techniques to handle manipulation of 3D as well as 2D objects. The authors begin by pointing out that Rotate-Scale-Translate (RST) transformations have become the de facto standard for 2D single and multi-touch interfaces. Preserving this standard for 2D applications, they want to extend it to 3D in a way that feels as natural as RST.
According to the authors, RST is popular because it is an intuitive 1-to-1 mapping of screen-points to users' fingers. It is a way to directly manipulate objects on the screen. But this currently does not exist for 3D objects.
The authors designed and developed a system that captures the 2D semantics of RST, but allows for any number of control points. The system works by mapping the points in an object's local space to point in screen-space using a continually updating transformation. The transformations are solved using a best-fit numerical approach so that the solution chosen is the one that minimizes the energy in the system. Using a best-fit solution ensures that adding extra input points does not significantly skew the results.
They then tested their own system to discover issues and try to resolve them.
Some of the problems encountered were in:
- controlling the two new DOF(degrees of freedom) which was resolved by learning the best way to rotate objects using both hands.
- resolving ambiguous rotational problems which occurs when the object rotates in the opposite direction from what is expected.
- resolving rotational exhaustion which occurs when the distance between two contact points change as the object is moved by the user.
- resolving error distribution
Discussion
This was a technical, and detailed paper which I had a hard time following at times,especially the parts where math formulas were involved. Nonetheless, I do think the solution presented by the authors is a good solution. Given the constraints they have, the solution given seems to be very realizable.
Yeah, I got lost in the math formulas too. Luckily they had a lot of pictures. Otherwise I would have gotten lost in some of their descriptions of how they rotate objects too. Still, as you say, they did present a reasonable solution. I think they presented the overall material in an understandable manner: present challenge first, then describe the solution.
ReplyDelete