HCC: Small: Eye Movement in Stereoscopic Displays, Implications for Visualization

Clemson University, Clemson SC

Investigators

Abstract

In this project the PI will investigate human eye movement within stereoscopic displays, with the goal of improving the perceptual effectiveness of these displays for visualization. While numerous studies have looked at eye motion for conventional displays, there has been very little work on stereo displays. The challenge and the opportunity for eye tracking in stereoscopic displays is that both eyes can be tracked, yielding data on gaze position on the screen and eye vergence, which affords an estimate of where the subject is fixating in three dimensional space (not just in screen space). Such data would shed light on the strategies users employ to explore and interpret information in these displays, which in turn may enable display designers to increase the effectiveness of their displays in transmitting information. This research will include technology development, perceptual experimentation, and visualization system development phases. The PI will initially devise an experimental methodology for binocular eye tracking devices, develop protocols and software, and run preliminary experiments to collect a baseline set of stereo tracking data which establish what tracking precision is possible and provide a basic understanding of how humans perform eye movements on stereo displays. Later, the focus will be on experiments exploring issues in the design of perceptually optimal visualizations on stereoscopic displays; eye tracking will illuminate how users scan such displays, providing insight into the visual strategies used to navigate, interpret, and extract information. Finally, the PI will employ what he has learned to design, construct, and evaluate an eye-tracking enhanced stereo 3D visualization system. Broader Impacts: The potential impacts of this project are as broad as the possible application of stereoscopic displays. For example, the visualization of geographic data often requires the display of layered surfaces; studies have shown that when using a stereoscopic display such visualizations can be designed so that the user's perception of the shapes of two overlapping surface layers is well preserved, but we lack empirical stereo eye tracking data at present to inform and guide this design. In the medical domain, stereoscopic displays have been shown to significantly improve endoscopic performance, in terms of laparoscopic precision, of both novice and experienced surgeons, but surgeons generally need to maintain a fixed viewing position in order to maintain stereoscopic integrity, a problem which could be alleviated by measuring positional and vergence eye movements and using them to dynamically adjust the location of the autostereoscopic "sweet spot." This work will contribute to building a perceptual science underlying stereoscopic display technology, which will be critical to its continuing perfection and application in domains such as science, education and training, healthcare, industry, and consumer products.