基于直方图比率特征的虹膜图像分类方法

Iris Texture Classification Based on Histogram Ratio Features

针对虹膜纹理的模式分类问题,提出了一种直方图比率特征的虹膜纹理分类方法。该方法利用虹膜图像的直方图信息,提取虹膜灰度等级对,通过灰度等级对自相关策略,计算虹膜纹理的直方图比率特征,实现了虹膜纹理的最大化分类。在相同的实验条件下,对不同样本的虹膜图像进行了仿真实验,结果表明:直方图比率方法较传统的直方图方法平均提高了3.05%的识别率。

Histogram is a widely used and important tool in image analysis and processing. Using histogram ratio features, an iris texture classification method is presented by analyzing characteristics of iris textures. Combining pairs of bins and computing corresponding count ratios, iris ratio features, given iris texture in an auto-correlative sense, is created to realize maximum classification of iris textures. Under the same experiment conditions, the proposed method and conventional histogram method are employed in different numbers of iris images, whose results show that histogram ratio features method improves 3.05% in correct classification rates for iris texture images compared with the conventional histogram method.