基于改进随机Hough变换的虹膜定位算法  

Iris Location Using Improved Randomized Hough Transform

作　　者：胡正平[1] 王成儒[1] 于莉娜[1] 

机构地区：[1]燕山大学通信电子工程系,秦皇岛066004

出　　处：《仪器仪表学报》2003年第5期477-479,489,共4页Chinese Journal of Scientific Instrument

摘　　要：针对随机 Hough变换 (RHT)中随机采样造成的大量无效采样和无效积累问题 ,提出了基于连续曲线段的 RHT(CRHT)算法并应用于虹膜定位中 ,它一方面利用了边缘图像的连续特性减少无效采样的数量 ,另一方面利用了精度不高的梯度方向信息减少无效积累的问题。实验结果证明改进的 CRHT算法提高了虹膜定位速度。To solve the problems of large amount of invalid samplings and useless accumulations when randomized Hough transform (RHT) is used to iris location, the continue curves feature based randomized Hough transform (CRHT) is presented and used to realize iris location. The method not only make good use of continue curve feature to reduce invalid sampling but also use the gradient direction information to reduce useless accumulations.Experiments show that the improved CRHT algorithms gained higher speed.

关 键 词：虹膜定位 连续曲线段随机霍夫变换 随机HOUGH变换 梯度方向信息 生物识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]