Thursday, March 3, 2011

Paper Reading #14, "PhotoelasticTouch"

http://stuartjchi.blogspot.com/2011/03/paper-reading-13-cosaliency-where.html
http://chi2010-cskach.blogspot.com/2011/03/paper-reading-10-bonfire-nomadic-system.html

PhotoelasticTouch: Transparent Rubbery Tangible Interface using an LCD and Photoelasticity
Toshiki Sato, Haruko Mamiya, Hideki Koike
University of Electro Communications, Tokyo
Kentaro Fukuchi
Japan Science and Technology Agency
Venue of presentation not specified in paper; presumably UIST.

Summary
PhotoelasticTouch  is designed to overcome the current limitations of tangible interfaces. The idea is for a tabletop system to be able to utilize an overhead table and an LCD screen to recognize deformations of an elastic object held in the user's hands. Any 3D elastic object can be used, either one provided or one created by the user (In this case, the researchers polyethylene and silicon rubber, to get the ideal balance between rigidity and deformability). Its position is detected by the camera detecting refraction of light from the table, and then displayed on the screen. Deformations are easily picked up and measured by the effect the stressed elastic has on the polarized light; the process is related to that used in photoelectric stress analysis.

Example applications given were a pressure sensitive touch panel (which I infer from the pictures to mimic the operation of a traditional touch panel by placing a sheet over the whole table, although use of more than one finger at once was certainly possible here) with which users were successfully able to manipulate a 3D object by pressure only, without sliding their fingers. The second application used an elastic face to allow interaction with a digital one. This wasn't much elaborated on.

The third and last was most fascinating. Taking a small piece of the elastic material, the user waves it in the air above the table, and when they pinch it the shape's shadow appears on the screen below. This enable the user of any shape provided as a paintbrush of that shape.

Some mentioned drawbacks of the system were the camera having difficulty functioning over a dark screen (as it runs on the polarized light from the table), occlusion of the camera by hands, and usage negatively impacting the transparency of the material. The light issue and the deteriorating transparency the researchers believe will be easily corrected.

Discussion
This project is pretty novel and has a lot of good ideas, although as the flaws section points out this was still just a proof of concept in a lot of ways. The best potential for applications here is for those that require specialized controllers; the transparent materials used here have to be less expensive than directly specialized electronics. I don't think I would want this for my base interface, but it can certainly be set to complement the mouse & keyboard rather than compete with them.

The biggest weakness of this research is the hands getting in the way of the camera. I realize this was already mentioned above, but it really is going to be the most difficult to solve and seriously impacts both degrees of freedom and precision of the input.

The paper itself was well done, I thought. It was well organized, readable, not given to excessive hyperbole, and unlike some previous UIST papers discussed applications and not just hardware.

The next direction to take this research should be, in addition to those the researchers mentioned themselves, trying to find some kind of material that meets the deformation requirements but is easily reconfigurable. That would massively increase the versatility of this platform.

Source: Paper

No comments:

Post a Comment