基于感兴趣区的虹膜定位算法

IRIS LOCATION ALGORITHM BASED ON REGION OF INTEREST

快速虹膜定位是实现虹膜自动识别系统的基础。利用小波变换后LL频带是图像内容的抽样缩略的性质,只对LL频带的图像进行虹膜定位,基于图像光斑特性选择定位感兴趣区域,用Hough变换定位虹膜内圆,单行梯度最大值方法定位虹膜外圆,然后按比例还原所有的参数值。试验结果表明这些参数与在原始图像中进行虹膜定位得到的参数一致,定位时间大大缩短,说明了该定位算法的有效性。

Foundation of realizing automatic iris recognition system is to locate iris quickly. After wavelet transformation, LL frequency band is the abbreviatory sampling of the image content, this property is utilized to localize the iris of image on LL frequency band only, the interested region is selected and located based on image facula characteristic, inner edge of the iris is located by Hough transformation, the outer circle of the iris is located by means of uniline gradient maximum value, then all parameters are restored pro rata. The experiment result proves that these parameters accords with the parameters obtained from the original image when localizing iris, the localization time is short-ened greatly, it shows that the algorithm is valid.