An Accurate Pupil Features Extraction Algorithm for Eye Tracking System
Determining the pupil center is fundamental for calculating eye orientation in Videooculography (VOG) systems. However, it is di眂ult to obtain a pupil center stably due to the effect of drooping eyelids, eyelashes, corneal reflection, and blinking. To solve this problem, we presented a novel method for determining pupil center by using a 3D transformation algorithm. It is assumed that the shape of the human eyeball is a pure sphere, and the contour of the pupil is a circle which belongs to a smaller virtual sphere concentric with the eyeball. Employing the anatomic knowledge and a single eye picture, a 3D eyeball model is reconstructed. And then the least squares plane fitting and circle fitting techniques are adopted to extract the whole edge points. Finally, the center of the pupil could be calculated accurately. Experiment results show the good performance of the algorithm.
Dongsheng Zhou Xiaopeng Wei Qiang Zhang
School of Mechanical Engineering Dalian University of Technology, China Dalian, 116024 Liaoning Key Lab of Intelligent Information Processing Dalian University, Dalian, 116622, China
国际会议
南宁
英文
2007-07-20（万方平台首次上网日期，不代表论文的发表时间）