Paper Reading #1: Interactive Interfaces: Spatial Interaction with Empty Hands and without Visual Feedback

Reference Information

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

By Sean Gustafson, Daniel Bierwirth, and Patrick Baudisch

Presented at UIST'10, October 3-6, New York, New York, USA

Author Bios

• Sean Gustafson is a PhD student working in Hasso Plattner Institute's Human-Computer Interaction Lab. He focuses on new forms of interacting with computers and holds bachelor and master's degrees in Computer Science from the University of Manitoba.
• Daniel Bierwirth received a master's degree in IT-Systems Engineering from Hasso-Plattner Institute and now works for Matt Hatting & Company UG and Agentur Richard GbR. He focuses on mobile application development and design thinking.
• Dr. Patrick Baudisch is the head of the Human Computer Interaction group at the Hasso-Plattner Institute. His research focuses on interaction techniques, especially with small and large devices.

Summary

Hypothesis

Can users interact successfully with an imaginary user interface without visual feedback and to what extent?

Methods

In three user studies, users consisting of young adults created simple drawings, edited existing ones, experienced deliberate interruptions, and pointed to locations in space based on a user-defined origin. An optical tracker installation tracked gestures based on markers on participants' gloves to minimize system limitations.

Results

Users were partially successful in using the system. The test of simple sketches and single stroke characters had a 95% success rate, which was greater than the target success rate. Users also increased in accuracy with repeated shapes. Multi-segment drawings did not fare as well, with faulty starting points leading to misalignment and condensed whitespace. The test of users' ability to remember where objects were even after a brief interruption from the task somewhat succeeded, with non-rotating participants and a reference point aiding in accuracy. The test of pointing at certain Euclidean coordinate points found increased error as the points moved farther from the user's non-dominant hand location.

Contents

Users worked with objects in a 2D space. Their non-dominant hand, producing an L shape, created the origin point for the plane. Spatial references are derived from the position of this hand. The system requires no devices to be held in-hand. This technology extends wearable computing, gestural input, mobile computing, and spatial interaction. The 3D coordinates created by the test were converted to 2D from the user's perspective. The first test presented users with a page of simple shapes to draw using their left hand as the origin and finished at user discretion. Some of the shapes were letters that posed problems with connecting strokes in prior studies. Others were repeated simple shapes and multi-stroke drawings. The multi-stroke gestures were not formally analyzed. This test suggested that users augmented their visuospatial memory throughout the test. The second test considered the longevity visuospatial memory after interrupting the participants. Initially, participants rotated 90 degrees after drawing a shape and indicated the requested corner of that shape. Notably, the left hand as a reference point increased accuracy even with an interruption. The third test considered accuracy on a coordinate plane, using the left hand forming the x and y unit vectors.

Discussion

This paper managed to measure user ability to use a screen-less, gesture-based interface. Their results suggest that this interface would work at least somewhat well for basic tasks. I am convinced that the authors of this paper had more success with their work than those against whom they compared the percentage of successes. This work is interesting because most computing today assumes the use of a screen. Taking away the visual feedback element poses an unusual question of whether we as people are capable of using the device. While this paper suggests that screen-less interfaces may be able to work successfully, I am not convinced that this is a viable solution for anything more than basic tasks. While this technology seeks to eliminate the problem of having to memorize many cumbersome gestures, the fact remains that a large body of gestures, some relatively unintuitive, must still be committed to memory for any tasks more complex than simply drawing a shape.