Influence of Depth Cues on Eye Tracking Depth Measurement in Augmented Reality Using the Magic Leap Device

Abstract

The eye-tracking technology is currently implemented in many mixed reality devices. However, eye-tracking measurements must be precise enough to permit correct localization in the environment; to allow, for example, linking such spatial positions to virtual objects. According to the human vision, major measurement issues would more likely occur in the depth axis rather than in the horizontal and vertical axes. In the literature, depth cues are known for improving human depth perception. In this study, our hypothesis is that, in an augmented reality environment, the more realistic virtual objects are displayed, thanks to depth cues, the more precise the eye-tracking device depth measures would be. Thus, using the MagicLeap device, we studied the effects of lighting and textures on eye-tracking depth measurement precision, by comparing the measures obtained under varying conditions of lights and textures, on both real and virtual objects. The results confirmed our general hypothesis, and we noticed a more significant influence of lights rather than textures on the precision of the measures. Moreover, we found that these depth cues reduce the measurement imprecision among observers, making the eye-tracking system more accurate when measuring depth.