As a discipline eye tracking is not something new. Already in the 19^th century people were experimenting with eye tracking. By simple observation it was already then possible to consistently follow the eyes of a subject in order to establish roughly the direction and characterize the movements.Later film recording of eye movements was introduced, so it was possible to post process the recordings, however still a manual task.With the more recent introduction of computers and digital cameras eye tracking has become fully automated, accurate and reliable.

The most reliable and widely used technique for obtaining eye tracking data is to illuminate the subject with infrared light and record using a camera that is sensitive in the infrared part of the electromagnetic spectrum. The infrared light ensures that the contrast between the pupil and the iris is optimal and allows easy and reliable detection. The images recorded from camera are fed to a computer which applies a series of sophisticated signal processing techniques to transform the images of the eyes into coordinates on the monitor (or environment).

In order to obtain high accuracy, the subject is required to be calibrated. This is a simple process where the computer registers how the eyes appear when the subject look at a number of predefined points on the screen (typically nine points).

Typical operating speed for an eye tracker is 30-60Hz (30-60 times per second). This range is adequate for most purposes. For special kinds of research special eye trackers are used that can go up to around 1000Hz.

The processing and physical foundation, of course, is somewhat more complex than here described. To get into more details you can check Wikipedia: http://en.wikipedia.org/wiki/Eye_tracking or the classic text book on eye tracking: A. Duchowski “Eye Tracking Methodology”.

An eye tracker delivers a wealth of data. Of primary interest is of course the gaze coordinates, from which the researcher knows where on the screen the subject was looking over time. These gaze data are typically superimposed on the images (stimuli) that were shown on the screen.

Other types of time dependent signals can be: distance, Fixation, Instrument, IR, camera, headbox, Hz, etc…