【摘要】 基于眼睛图像的几何特性,提出了一种基于最长弦的瞳孔定位算法,该算法能快速地检测到瞳孔的位置。由于图像中的瞳孔并不是规整的圆形,此方法与传统的基于圆灰度梯度检测算子的瞳孔定位算法相比能够更好地拟合瞳孔边界。在精确地定位瞳孔之后,根据瞳孔和虹膜近似同心的特性,利用圆灰度梯度检测算子在小范围内搜索虹膜的外边沿。与传统的虹膜定位方法相比,此算法避免了搜索的盲目性,提高了虹膜定位的速度和精度。更多还原

【Abstract】 This paper proposes a new method for pupil detecting based on the geometrical characteristic of eyes,which can detect the pupil rapidly.Pupils in captured images are not regular circles,the longest-chord based method can de-tect the boundary of pupil more effectively than those conventional methods based on circular template.The pupil and the iris are approximately concentric circles,we can search the outer boundary of the iris in less range with the circu-lar template.By contrast with the usual method based on circular template,this method reduces the blindness of the search process,and improves the speed and accuracy.更多还原