The eye’s reflection becomes a signal
Most eye trackers use near-infrared light, which people generally cannot see. A small camera records the eye as this light reflects from the curved, clear surface at the front of the eye—the cornea. The reflection appears as a bright point, often called a corneal reflection.
At the same time, the camera locates the pupil, the dark opening at the center of the eye. As the eye turns, the pupil shifts position relative to the reflected point. Tracking both gives the device a more dependable clue to the eye’s rotation than following the pupil alone, which can move in the camera’s view as the head shifts.
From eye movement to a point on a screen
The tracker combines those measurements with a calibration: the person looks at known spots, allowing the system to learn how particular eye positions correspond to locations in the device’s view. It can then estimate where the person is looking, whether that means a place on a monitor or a point in a room.
This is an estimate, not a direct reading of attention or intent. The system identifies the direction of the eyes and maps it to a location; it cannot tell from that signal alone what the person is thinking about. Head movement, glasses, eyelids, lighting, and small differences between people can also affect the measurement, so devices may need calibration or adjustment.
The key is the changing geometry between pupil and reflection: a tiny shift in that relationship lets a camera turn an otherwise subtle eye movement into a usable location.
