基于Gabor滤波器组的虹膜特征提取与匹配

Iris Feature Extraction and Matching Based on Gabor Filter Groups

提出了一种基于Gabor滤波器组的虹膜识别方法,能有效地从虹膜图像中确定一个感兴趣的区域(ROI)并用于多尺寸多方向地提取虹膜图案的空间和方向特征,然后通过寻找相关特征矢量之间的最小距离实现虹膜图案匹配.实验结果证明我们的方法既能对虹膜图案的比例、平移保持不变性,又能消除或减少因睫毛、眼睑和反光造成的干扰.

This paper presents an iris recognition method based on Gabor filter groups, which effectively extracts the spatial and directional features of a region of interest （ROI） of iris patterns in multiple scales multiple directional. Then, iris pattern matching is performed based on finding the distance between the corresponding feature vectors. Experimental results demonstrate that the proposed method not only keeps the invariability both in scale and translation of iris patterns, but also effectively eliminates or decreases the interference of eyelash, eyelid, and reflected light.