This paper proposes a fast and robust algorithm for detecting the center of pupil which is fundamental of eye movement tracking systems. The algorithm first extracts an eye region by segmenting the image captured by a visual CCD camera. Recognizing processes are then applied to the eye region and sample points at the edge of pupil is detected. Finally, double circle fitting is done for the edge set and the center of pupil is calculated. This method has been tested by ten people with both naked and glasses-wearing eyes under different lighting conditions. It proved the robustness and accuracy of the algorithm to wide variety of images.
展开▼