Eye pupil detection using the least squares technique

The paper presents a technique used for the detection of the eye pupil center coordinates by using the least squares approach. The algorithm realizes the fitting of the ellipse to a group of input data points provided by the eye image collected using a video camera in infrared illumination. The method can be used in the area of assistive technology to implement an eye tracking system used to communicate with neuromotor disabled patients. The system has been tested on a set of healthy individuals in laboratory conditions. The tests have shown higher accuracy, lower noise and lower algorithm running time as compared to other similar techniques presented in the literature.