Online 3D Gaze Localization on Stereoscopic Displays

Abstract

This article summarizes our previous work on developing an online system to allow the estimation of 3D gaze depth using eye tracking in a stereoscopic environment. We report on recent extensions allowing us to report the full 3D gaze position. Our system employs a 3D calibration process that determines the parameters of a mapping from a naive depth estimate, based simply on triangulation, to a refined 3D gaze point estimate tuned to a particular user. We show that our system is an improvement on the geometry-based 3D gaze estimation returned by a proprietary algorithm provided with our tracker. We also compare our approach with that of the Parameterized Self-Organizing Map (PSOM) method, due to Essig and colleagues, which also individually calibrates to each user. We argue that our method is superior in speed and ease of calibration, is easier to implement, and does not require an iterative solver to produce a gaze position, thus guaranteeing computation at the rate of tracker acquisition. In addition, we report on a user study that indicates that, compared with PSOM, our method more accurately estimates gaze depth, and is nearly as accurate in estimating horizontal and vertical position. Results are verified on two different 4D eye tracking systems, a high accuracy Wheatstone haploscope and a medium accuracy active stereo display. Thus, it is the recommended method for applications that primarily require gaze depth information, while its ease of use makes it suitable for many applications requiring full 3D gaze position.

Comparing estimated gaze depth in virtual and physical environments

“…The following pairwise comparisons with Bonferroni correction revealed significantly different errors between depths of 5 and 15, -5 and 15, and between 15 and -15 cm. Results generally corroborate previous findings of signed vergence error on stereoscopic displays [Wang et al 2013].…”

“…Gaze depth, i.e., vergence, can be measured via spatial triangulation [Pfeiffer 2010]. The geometry of the approach is given by Wang et al [2013]. If a binocular eye tracker delivers two on-screen gaze points, (x l , y l ) for the left eye and (xr, yr) for the right, then the horizontal disparity ∆x = xr −x l is sufficient to estimate gaze depth z [Duchowski et al 2011].…”

“…Inter-ocular distance α can be assumed to be constant at 6.3 cm [Smith and Atchison 1997]. Although eye separation varies across individuals [Dodgson 2004], and significantly affects spatial triangulation, for a fixed-screen display a 3D calibration procedure is used to accommodate for this variation [Wang et al 2013] (see below).…”

A new approach to create 3D fixation density maps for stereoscopic images

“…Wang et al had studied this problem and proposed a 3D calibration method [6]. This method was adopted with little changes to fit to our actual 3D calibration situation.…”

“…They proposed a method to measure gaze depth in [3]. After that, a method to improve accuracy of eye-tracking data under stereo viewing was proposed in which they followed the eyetracker's protogenetic 2D calibration with a 3D calibration [6]. Another approach called Parametrized Self-Organizing Map (PSOM) proposed by Essig et al [7] was designed to solve the 3D calibration problem by mapping two 2D screen coordinates to a 3D gaze point.…”