|
Eye
Tracking is the methodology of measuring and recording the eye
movements relative to the head position of an observer or that of
capturing the gaze on a visual scene. The meaning of visual scene is
wide and it is possible to be related to an analog product (for example
an analog map) or to a digital product (for example a digital map) that
is depicted on a computer monitor or on a television display or
projected with an appropriate device to a flat surface. Also, it is
possible that the natural surface might be the referred visual scene.
Several techniques have been used for the detection of eye movements.
These methods are based on the generation of an electromagnetic field
that is placed around the eye (electro-oculography) or the use of a
special type of contact lenses. However, the technique that has been
dominated, mainly due to its direct application, is based on the use of
devices that are able to record the gaze on the visual scene analyzing
eye images. At the Cartography Laboratory of N.T.U.A. the Viewpoint Eye
Tracker® by Arrington Research is installed, whose operation is based
on the latest technique.
|
|
The geometry of the system operator – observer - EyeTracker |
The
Eye Tracker is a tool for the tracking of gaze observer during the
observation of visual scenes on a computer monitor. The basic
components of the system are: the recording device-hardware and the
processing software. The Eye Tracker is embedded in a computer machine,
which has an additional card for graphics. This contributes to the
support of two monitors; the first one is used by the tracker's
operator (primary monitor) and the second is used to project visual
scenes. The recording device in the central processing unit (CPU)
consists of a board connected on the motherboard and externally of a
system attached to the observer's optical system. The system of
attachment supports the existence of a camera and an infrared LED for
each eye of optical system. The system's geometry is integrated by a
mechanism that stabilizes the position of the optical system of the
observer. The Eye Tracker also provides the opportunity of “extending”
visual scene to physical surfaces, as it includes a third camera placed
on the system attached to the observer's head. |
|
The geometry of the system camera – infrared LED – mechanism for head stabilization
|
The
system’s function is accomplished by a set of sequential operations.
The infrared LED, which is located under the observer's eye,
illuminates the eyeball while its image is being capturing by the
camera that is placed next to the infrared LED. The illumination of the
infrared LED allows the discrimination between regions of pupil and
iris. Furthermore, the camera also captures the corneal reflection by
applying a predefined threshold. The center of the pupil and the
position of corneal reflection are detected by using segmentation
algorithms. For each observer, the system creates a
transformation that corresponds to every movement of central vision and
the relative movement on the stimuli that is depicted on the secondary
monitor. The generation of this function occurs through a process of
auto-calibration, which takes place for each observer. After defining
the appropriate function, the Tracker is able to detect each movement
on stimuli scene or each eye movement that occurs on the natural
surface. |
|
Diagram with the system's operations |
The
Tracker uses three different methods to detect the visual center. The
recording system detects the center of vision by locating the center of
the pupil (pupil location method) or locating the position of the
corneal reflection (glint location method) or computing the vector
between pupil and the position of corneal reflection (pupil-glint
vector). The process of auto-calibration takes place by the
observation of fixed point-targets that are depicted as squares on the
visual scene. This process has several variables, for example the total
number of the fixed targets and the time that they are represented,
which are selected by the requirements of the experimental design. The
recorded data do not contain only the records of gaze coordinates on
the visual stimuli. The Tracker provides the ability to create special
regions of interest (ROIs), to capture the range of observer's pupil,
to compute the ratio of the tangents pupil, to compute the eye torsion
and to record the fixation. Furthermore, each record is followed by a
pointer of its quality. The sampling frequency is being selected by the
operator, who has to select among two available choices (30Hz or 60Hz).
The visual range is ±44 degrees of the visual arc horizontally and ±20
degrees in the vertical dimension. The accuracy of the EyeTracker is in
the range between 0.25-1.00 degrees of the visual arc, when the
recording resolution is 0.15 degrees of the visual arc. |
|
The detection of the center of pupil (left image) and the detection of pupil and corneal reflection (right image) |
The
EyeTracker is a useful system for implementing cartographic experiments
or similar experiments related to the process of the visual search of
psychological studies, as well as, for advertisements or computer
software evaluation. The system opens innovative experimental
ways to support research on cartography, because it provides recordings
of the image of visual trace while capturing the time of the
observation, that indicates the “way” and the fixation points that
occur during the observation of a scene, a stimuli or an event. |
|
The visual trace during the search of a symbol on a cartographic background |
Bibliography
Arrington Research, 2005, ViewPoint EyeTracker. PC-60 Software User Guide, Arrington Research Inc, Scottsdale (U.S.A.).
Duchowski A., 2007, Eye Tracking Methodology. (2nd edition), Springer-Verlag, London.
Krassanakis V., 2009, Recording the trace of visual search: A research method of the selectivity of hole as a basic shape characteristic. Diploma Thesis, School of Rural & Surveying Engineering, National Technical University of Athens, Athens (In Greek).
|
Edited by
Vassilios Krassanakis
Diploma of Rural & Surveying Engineering of N.T.U.A.
PhD Candidate
krasvas@mail.ntua.gr
|
|