IEEE Consumer Electronics Magazine - October 2015 - 70

attention or intent, invalidating the results. Long fixations could
indicate either attraction to an element or difficulty in understanding it. It also requires specialized and expensive setup.

gAze trAcking in virtuAl
And Augmented reAlity
Virtual characters and environments have crossed the realm of
gaming [2] to find use in interactive applications such as virtual tourism, urban planning, and online education (Figure 5). To
enhance user experience in such advanced virtual reality (VR)
applications, user gaze is actively tracked in real time to bring
scenes into sharp focus by determining the user point of view
in three-dimensional space, facilitate user navigation by deciphering visual attention, and controlling a virtual gamer's
responses according to speed of eye movements. Important
contributors to this field are Sony, Nintendo, and Sensor
Motoric Instruments (SMI) GmBH, whose gaze-coupled VR
applications extend from gaming to gaze interactions and ergonomics [21], [22].

overall, it can be stated that
despite decades of research in this
field, the realization of a winning
implementation that provides
both reliable and accurate
gaze-controlled interfaces in
the consumer domain is still
some distance in the future.
mAjor limitAtions
Key issues in eye tracking for VR and augmented reality
involve Midas touch, gaze-point tracker inaccuracies leading
to incorrect actions/selection, and uncertainties in tracking
resulting from miscalibration or latency.

eye gAze For security And AuthenticAtion
Enhanced resistance against shoulder-surfing attacks has
encouraged the development of gaze-based password schemes
[25]. While the simplest of such authentication methods include
entering alphanumeric characters as a security PIN with eye
gaze, advanced security options involve gaze-image passwords,
i.e., password entry by gazing with a specific sequence on an
image on a screen or at multiple features in an image [23].

mAjor limitAtions
The prime hindrances to such systems are the possibility of identity theft and that unintentional eye and body movements may
lead to erroneous gaze points and incorrect password entry.
70 IEEE Consumer Electronics Magazine

october 2015

CONCLUDING THOUGHTS
The potential of gaze-based interactions in modern consumer
devices and platforms is still being explored. Major challenges
exist due, in particular, to the highly dynamic modalities and
unconstrained use cases for handheld devices and in the automotive sector due to the rigorous safety considerations and the
wide range of lighting conditions experienced in the vehicle
cockpit. The requirements for a consumer-grade gaze-tracking
system include real-time, high-accuracy gaze estimation, easy
calibration (calibration-free is preferable), and robust responses to widely varied lighting conditions. Robustness to head
pose is also important, although this is a more tractable problem if a high-accuracy head model is available. As in all consumer electronics system designs, it is essential to maximize
system performance while reducing unit costs.
From this point of view, two of the most powerful methods for gaze estimation developed to date are presented in
this review, appearance-based and feature-based techniques.
While each method has its respective advantages and drawbacks, the appearance-based models, because of their lower
hardware requirements, would appear to offer a more feasible solution for mobile platforms. However, they do suffer
from lower accuracy and are sensitive to changes in head
pose, illumination, and facial expression. Alternatively, feature-based or corneal-glint methods can provide significantly higher accuracies, but due to a requirement for complex
hardware, stable positioning, and detailed calibration, they
are really only suited for indoor and desktop environments.
Overall, it can be stated that despite decades of research in
this field, the realization of a winning implementation that provides both reliable and accurate gaze-controlled interfaces in the
consumer domain is still some distance in the future. In fact,
much of the research in this field only serves to emphasize that
the accurate and nonintrusive measurement of eye gaze in real
time and in unconstrained use cases using consumer-grade technology remains a major challenge.
More importantly, the ultimate prospect to employ eye gaze
as a means of HCI depends entirely on a very nontechnical factor, and one that is extremely difficult to quantify-the simplicity and ease-of-use of this sophisticated technology. Unless
it can be realized in a robust, reliable, and calibration-free
form, then consumers are unlikely to engage with and adopt
the technology. And without the magic of simplicity, it is
unlikely to become a sustainable success and a new and practical mode of HCI.

ABOUT THE AUTHORS
Shabab Bazrafkan (sbazrafkan@fotonation.com) is with the
Center for Cognitive, Connected, and Computational Imaging, National University of Ireland, Galway.
Anuradha Kar (akar@fotonation.com) is with the Center
for Cognitive, Connected, and Computational Imaging,
National University of Ireland, Galway.
Claudia Costache (ccostache@fotonation.com) is with
the Center for Cognitive, Connected, and Computational
Imaging, National University of Ireland, Galway.

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - October 2015