基于多结构鲁棒估计的虹膜外边缘定位方法  

Iris Location Method for Outer Edge Based on Robust Estimation for Multi-structure

作　　者：林绍辉[1] 陈水利[1] 吴云东[1] 

机构地区：[1]集美大学理学院,福建厦门361021

出　　处：《集美大学学报（自然科学版）》2014年第1期63-68,共6页Journal of Jimei University：Natural Science

基　　金：福建省自然科学基金资助项目(2011J01013);福建省科技厅产学研重大项目(2011H6020);厦门市科技基金项目(3502Z20110010;3502Z20123022)

摘　　要：在Adaboost检测人脸区域的基础上,提出了基于多结构鲁棒估计的虹膜外边缘定位方法.通过对200幅单人脸的图像数据集的仿真实验,结果表明,本方法相比于传统的RANSAC模型生成算法以及Hough变换等方法,能得到更精确的定位结果,而且能加快有效的模型生成,虹膜外边缘定位精度达到94%.On the basic of detecting the area of human face with Adaboost algorithm, a iris location method for the outer edge based on robust estimation for multi-structure was proposed. For 200 single human face image data set, the simulation experiments showed that more accurate location was obtained and the valid hypothesis generation was accelerated by the proposed method which compared with the conventional hypothesis generation method of RANSAC and Hough Transform. In addition, the accuracy of iris location for outer edge achieves 94%.

关 键 词：鲁棒估计 多结构 相似函数 ADABOOST CANNY边缘检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]