A new algorithm for quantitative analysis of eye movements recorded with an infrared CCD camera

We report a new algorithm that allows quantitative analysis of eye movements recorded with an infrared CCD camera, now widely used to examine nystagmus. With this camera, it is possible to observe minute eye movements and the recorded eye movements can also be replicated. We first divide a segment of a film of eye movements into frames and then subdivided the area of each frame into multiple blocks. A pattern-matching method is used to determine the direction and distance (i.e., vector) of the movement of the pixels in each block. The axis and angle of rotation of the eyeball are then calculated based on the derived vectors. Our algorithm has several advantages. It does not assume that the eyeball rotates around the center of the pupil, but calculates the axis of rotation. Accordingly, complicated eye movements can be analyzed, e.g., those in which the axis of cyclorotation does not cross the center of the pupil. It also minimizes the effects of noise caused by the eyelids, because multiple vectors are used to determine the eye movements. Therefore, it facilitates analysis of eye movements in patients with narrow palpebral fissures. It also eliminates the effects of infrared ray reflex on the bulbar conjunctiva.