HCI Research Papers: Paper Reading #1: Imaginary Interfaces: Spatial Interaction with Empty Hands and without Visual Feedback

Imaginary Interfaces: Spatial Interaction with Empty Hands and without Visual Feedback

Sean Gustafson, Daniel Bierwirth, and Patrick Baudisch. 2010. Imaginary interfaces: spatial interaction with empty hands and without visual feedback. In Proceedigns of the 23rd annual ACM symposium on User interface software and technology (UIST '10). ACM, New York, NY, USA, 3-12. DOI=10.1145/1866029.1866033 http://doi.acm.org/10.1145/1866029.1866033
UIST 2010 New York, New York.

Sean Gustafson is a PhD student at the Hasso Plattner Institute located in Potsdam, Germany. After receiving both his Bachelor and Master's degrees in Canada he worked as an embedded systems developer of medical devices.
Daniel Bierwirth received his Bachelor degree in Computer Science and Media from Bauhaus University in Weimar, Germany in 2008 followed by his Master's degree in IT-Systems Engineering from the Hasso-Plattner Institute in Potsdam, Germany in 2011. He cofounded both Matt Halting & Company UG as well as the Agentur Richard GbR.
Patrick Baudisch earned his PhD in Computer Science from Darmstadt University of Technology in Germany. Prior to becoming a professor at Hasso Plattner Institute, Baudisch reasearched adaptive systems and interactions at both Microsoft Research and Xerox PARC.

A screen-less input device, consisting of an imaginary plane, is feasible for practical use.
Participants would perform fewer Graffiti recognition errors than were previously reported by Ni And Baudisch by utilizing their visuospatial memory.
Intermittent activities will affect visual memory and therefore reduce accuracy.
Utilizing one hand as a frame of reference will increase accuracy.

For the three accuracy studies performed the researchers compensated groups of 12 people, mainly in their twenties, to come perform some tests. The testing environment was a closed room with an optical tracking system consisting of eight cameras set up. The participants wore tracking gloves. In a couple of the studies additional equipment was added, such as a foot pedal. The information collected through the tracking system was rotated and transformed into two-dimensional data as the imaginary interfaces system is based on the idea of interacting with an invisible plane in front of the user. A computer voice was utilized in one test to allow for a random ordering of coordinate requests. The participants were instructed about what they would be doing prior to beginning each session.

Although the study group was not large enough to gather statistical data, the information that was collected from the first study appears to support the hypothesis that fewer Graffiti recognition errors were realized utilizing this system than previous systems. The failure rate of the recognition in this study was a mere 5.5% as compared to 15% in a system that doesn't contain a reference point. Additionally, the claims that visuospatial memory fades both over time and with intermittent activities appears to be supported as well. The first of those two claims was tested when participants drew the same shape five consecutive times and the difference between the accuracy in the first and the last shape was much greater than the difference between the last two shapes. When participants had to turn 90 degrees between activities, their accuracy decreased aiding the idea that other activities fade visuospatial memory. Additionally, accuracy when pointing increased as the distance to the reference point decreased. Finally, the accuracy testing performed was done in a controlled room but a prototype for a wearable device was created and tested.

A looming problem in mobile computing systems is the size limitation imposed by the visual feedback currently given through screens. These screens can only shrink to a certain size become they lose their functionality, either from displayed information being too small to read or interactions becoming too inaccurate. The authors explore other possibilities, by allowing input without a visual feedback element. Further research is suggested to add audible cues and extend the project into the third dimension. The concept was shown to be sound which will have implications in the future of miniaturization.

This is very exciting new research, however maybe not exactly as the authors had intended. I cannot see this research creating an input system that does not have any visual feedback yet is still accurate enough to chart stocks. The authors did indeed show that simple input without a visual aid is not only possible but plausible as well. The pointing, or button pressing, for example worked fairly well for brand new users to a completely different system and with a little bit of practice would certainly improve. However, without audio or visual clues, the user would appear to be limited to a predefined number of buttons that they have committed to memory. Although this is slightly addressed in the paper, during the discussion about improved accuracy for stock charting, I can see this system being used for simpler activities. For example, I hate going through the hassle of taking my phone out of my pocket to only change the song or check why it just vibrated. It would be much simpler to simply throw my hands in front of me, make an "L" shape with my left hand and push an imaginary button with my right hand to start a new song or have incoming information read to me. Some may see this as lazy, but it would in fact be a fantastic time saver, and even a slight increase in efficiency translates to real world benefits in the modern world.

Picture Source: "Imaginary Interfaces: Spatial Interaction with Empty Hands and without Visual Feedback"

HCI Research Papers