1. Introduction to Eye Tracking
Antti Aaltonen
Tampere University Computer Human Interaction Group

2. Content Physiology of Eye (1) Eye Movements (3) Vision (2)
Visual Perception (1)
Terms (1)
Eye Tracking Methods (6)
Video-based Eye Tracking Systems (2)

3. Physiology of Eye
Cornea is a transparent structure that covers the iris and pupil; a part of the focusing system of an eye.
Pupil is the adjustable opening at the center of the iris that allows varying amounts of light to enter the eye.
Lens helps to focus light on the retina.
Retina includes rods (94%), which are sensitive to light and cones (6%) that capture colors. Cones are concentrated in the centre of the retina - the fovea

4. Eye Movements Eyes move all the time (even during sleep)
Several different movement types, such as
Pursuit
Tremor
Rotation
Drift
But the most interesting types are
Fixation
Saccade

5. Fixation
Eye is a (relatively) still and “fixated” to the certain point. E.g. reading a single word.
All the information from the scene is (mainly) acquired during fixation.
Duration varies from ms, typically ms.
Typical fixation frequency is < 3 Hz
Interspersed with saccades...

6. Saccade “Jumps” which connect fixations
Very rapid -- duration is typically only ms
Very fast (up to 600 o/s) and therefore the vision system is suppressed during the movement
Ballistic; the end point of saccade cannot be changed during the movement
Saccades are used to move the fixation point
If larger than 30 degree movement is required, head moves along with eyes

7. Vision1 Vision field is divided in to three regions
Fovea provides the sharpest vision
Parafovea previews foveal information
Peripheral vision reacts to flashing objects and sudden movements
Peripheral vision has 15-50% of the acuity of the fovea and it is also less colour-sensitive

8. Vision2 0 1 5 Foveal Parafoveal Peripheral
In real life the regions are asymmetric
E.g., in reading so-called Perceptual span (size of the effective vision), is 3-4 letter spaces to the left of fixation and letter spaces to the right
1O of visual angle is roughly equivalent to 3-4 letter spaces
Parafoveal
Peripheral
Foveal
[deg]

9. Visual Perception
The perception of a complex scene involves a complicated pattern of fixations, where the eye is held (fairly) still, and saccades, where the eye moves to foveate a new part of the scene
The main issue is “Where to look next?” Answer may be that perception is influenced with
Cognition, e.g. “meaningful” and parts previewed with parafoveal vision
Visual cues; highly visually attractive areas attracted by peripheral vision

10. Eye Tracking Methods Rough taxonomy Electronic methods
Mechanical methods
Video-based methods
Single point
Two point

11. Electronic methods
The most used method is to place skin electrodes around the eyes and measure the potential differences in eye
Wide range -- poor accuracy
Better for relative than absolute eye movements
Mainly used in neurological diagnosis

12. Mechanical methods Based on contact lenses with Very accurate
mirror planes + reflecting IR-light
coil + magnetic field
Very accurate
Very uncomfortable for users who are not used to wear lenses
Usable only for lab studies

13. Single point video-based methods
Tracking one visible feature of the eyeball, e.g.:
limbus (bondary of sclera and iris)
pupil
A video camera observes one of the user's eyes
Image processing software analyzes the video image and traces the tracked feature
Based on calibration, the system determines where the user is currently looking
Head movements not allowed
Bite bar or head rest is needed

14. Two point video-based method1
The same idea as in the single point method except now two features of eye are tracked – typically
corneal reflection
pupil
Uses IR light (invisible to human eye) to
produce corneal reflection
cause bright or dark pupil, which helps the system to recognize pupil from video image
Bright pupil
Corneal reflection

15. Two point video-based methods2
The optics of the system can be mounted on
head
floor
If optics are floor mounted, the system is not in contact with the user
Generally head movements are not restricted and they can be separated from eye movements, but...
With floor mounted optics the system has to track the user’s head in order to keep the eye in the field of view of camera, which limits the head movements. The performance can be improved with
servo controlled tracking mirrors, or
a camera taking a wide-angled view of the user’s head and by using artificial neural network, the system searches the eye from the image.

16. Some terms Accuracy Spatial Resolution
The expected difference in degrees of visual angle between true eye position and mean computed eye position during a fixation.
Because of the vision system and physiology of eye the accuracy is usually 0.5-1O.
Spatial Resolution
The smallest change in eye position that can be measured.
Temporal Resolution (sampling rate)
Number of recorded eye positions per second.

17. ASL 4000-series (1996) Main components
Floor mounted optics
Control unit
2 computers (control & subject)
Head movements (partially) compensated with tracking mirrors and extended head movement options
Temporal resolution 50 Hz
Spatial resolution 0.5O
Tracks only one eye
Poor analysing software and no API

18. SMI EyeLink (1999) Contains Temporal resolution 250 Hz
Head mounted optics
2 computers (control & subject)
Temporal resolution 250 Hz
Spatial resolution <0.01O
Tracks both eyes
Reasonable analysis software
WIN API’s for Microsoft Visual C++