At least 50 years ago, someone consciously recognized that the direction of a person's gaze was directly related to the relative positions of the pupil and the reflection of an object off the cornea. In the late 1960's Kenneth Mason formalized the "pupil-center/corneal-reflection" method, an automated procedure for observing the eye with a camera, measuring the locations of the pupil center and corneal reflection, and calculating the direction of gaze.

In the early 1970's John Merchant and Richard Morrisette, in work sponsored by the U.S. Air Force, built a system that reduced the concept to practice. Their famous "Oculometer" used a video camera to observe the subject's eye and a computer to process the camera's image of the eye. Their image processing algorithms consisted of innovative methods to a) recognize the pupil of the eye and calculate its geometric center, and b) locate the relative position of the corneal reflection. They introduced the use of higher order polynomial equations to correct for nonlinearities in the Oculometer, and they developed root-mean-square regression methods for calibrating the equations to individual people's eyes.

The accuracy of eyetracking systems depends in large measure on how precisely the image processing algorithms can locate the relative positions of pupil center and the corneal reflection. Though it is possible to determine the boundary of the pupil in a normal picture of the eye, early eyetracking systems used the bright-eye effect to enhance the image of the pupil, significantly increasing the accuracy of pupil location. To achieve the bright-eye effect, light is shined into the eye along the axis of the camera lens. The eye's lens focuses the light that enters the pupil onto a point on the retina. Because the typical retina is highly reflective, a significant portion of that light emerges back through the pupil, and the eye's lens serendipitously directs that light back along the camera axis right into the camera. Thus the pupil appears to the camera as a bright disk, which contrasts very clearly with the surrounding iris.

To achieve the bright-eye effect, early eyetracking systems used a light source located to the side of the camera lens and a semi-silvered mirror, mounted in front of the lens at 45 degrees, to reflect the light along the camera axis into the eye. Though the method achieved the bright-eye effect, the semi-silvered mirror reflected away half the light coming back from the eye, reducing the brightness and clarity of the camera's image of the eye. In 1986 Thomas Hutchinson developed the idea of placing a small light-emitting diode (LED) in the center of the camera lens. The equipment configuration is simpler, and, because the LED blocks only a small percentage of the camera aperture, the image of the eye is brighter and clearer.

In 1988 LC Technologies introduced the first PC-based eyetracking system that interfaces easily with other equipment and computers. Dixon Cleveland developed advanced image processing algorithms for locating the pupil and corneal reflection more accurately and consistently. He also developed automatic focusing methods which maintain gazepoint prediction accuracy as a person moves his head toward and away from the camera, making the system more tolerant to head motion.

There are several other methods for measuring eye orientation. The electro-oculography method, for example, involves measuring electric potential differences between locations on different sides of the eye. Most of these methods, however, while being significantly less expensive, do not provide the accuracy of the pupil-center/corneal-reflection method. Cornsweet, on the other hand, developed the Purkinje method which uses a camera and light source and computes the eye's orientation based on light reflections from both the front and rear surfaces of the eye's lens. Because it does not depend on the pupil opening and closing concentrically about the eye's optic axis, the Purkinje method can be more accurate than the pupil-center/corneal-reflection method, but it requires a significantly more controlled lighting environment to be able to detect the rear surface reflection off the eye's lens.