三次B样条曲线拟合的虹膜定位  

Iris Localization Algorithm Based on Cubic B-spline Curve Fitting

作　　者：叶永强[1] 沈建新[1] 周啸[1] 张敏[2] 

机构地区：[1]南京航空航天大学机电学院,南京210016 [2]四川大学华西口腔医学院,成都610041

出　　处：《小型微型计算机系统》2011年第9期1885-1889,共5页Journal of Chinese Computer Systems

基　　金：教育部"新世纪优秀人才支持计划"项目(NCET-06-0502)资助;江苏省"333工程"项目(2007-16-59)资助

摘　　要：采用圆检测定位虹膜内外边界的方法是当前虹膜定位的主流算法.当虹膜图像分辨率很高时,圆曲线不能准确地拟合虹膜真实边界,特别是受瞳孔收缩影响很大的内边界.而采用三次B样条曲线能够很好地拟合内边界.为了提高定位效率,首先运用质心探测方法分割出瞳孔区域,然后在瞳孔区域中搜索内边界点,采用三次B样条曲线精确拟合内边界;最后利用Can-ny算子检测外边界,并采用圆曲线的最小二乘拟合外边界.运用Bath大学虹膜库中的1000幅虹膜图像对该定位算法进行测试,内边界定位时间0.0203s、准确率99.2%;外边界定位时间2.0277s,准确率98.9%,满足准确、高效的定位要求.The current important methods of iris localization are based on circle detection. But they could not fit the real boundary well when iris images are high-resolution, especially the inner boundary under the influence of pupil constriction. Proposed method based on cubic B-spline curve can figure out this problem. It locates the inner boundary area first, and then the outer boundary. To improve the efficiency and robustness for inner boundary localization, this paper has proposed a method to segment the pupil area first based cenavid detection, then search inner edge in the segmented area. The outer boundary area is then determined by using the parameter relations between inner and outer boundary. Finally, using canny operator detects the outer edge, and the outer boundary is fitted in the Least-square circle sense. The Experiment results based on the iris database of Bath University, with 99.2% accuracy and 0. 022s positioning time of inner boundary, 98.9% and 2.027s of outer show that the proposed approach is efficient and robust.

关 键 词：虹膜定位 质心探测 三次B样条曲线 最小二乘拟合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]