一种虹膜色素块检测与分类方法

针对现有虹膜识别系统中全局纹理特征提取方法忽略了纹理类型信息的问题,提出了一种针对全局性纹理中虹膜色素块的检测与分类方法.该方法利用灰度聚类法实现虹膜图像中色素块可能存在区域的初定位,依据坑洞和色素斑这两类色素块的灰度空间分布特性,定义一组区域特征参数作为分类特征向量,利用支持向量机实现二者的检测与分类.算法对图库中图像的坑洞和色素斑的检测正确率分别为99.2%和86.5%,对无特征纹理存在的虹膜图像检测正确率为87.2%.实验结果表明,该方法具有较高的正确率,能够满足虹膜识别系统的纹理特征提取要求.

A detection and classification method for iris pigment block

In order to solve the problem that the global texture feature extraction methods in the existing iris recognition system ignore the information of texture types, a detection and classification method for iris pigment blocks based on global texture was proposed. The initial location of probably existing regions of pigment blocks in the iris images was realized with a gray cluster method. According to the gray spatial distribution characteristics of two pigment blocks including plaques and crypts, a set of region feature parameters were defined as the classification feature vector. In addition, the detection and classification of iris plaques and crypts were realized by a support vector machine. The detection accuracy for crypts and plaques of images in the gallery are 99.2% and 86.5%, respectively. Moreover, the detection accuracy of iris images without any feature textures is 87.2%. The experimental results show that the proposed method has higher detection accuracy, and can meet the requirement of texture feature extraction in the iris recognition system.