基于单频2D-Gabor滤波器的虹膜识别算法研究

针对2D-Gabor滤波器设计的问题,提出了一种单频2D-Gabor滤波器设计方法。该方法在详细分析了2D-Gabor滤波器中各个参数意义的基础上,通过二分法的思想确定适合虹膜纹理的单一频率ω,再根据其它参数与其关系确定出各个参数取值,从而获得单一频率下的2D-Gabor滤波器。实验结果表明使用设计的单频2D-Gabor滤波器对虹膜图像进行特征提取可以使虹膜识别的识别率达到98.61%以上。使用单频2D-Gabor滤波器不仅可以提高特征提取速度,降低存储空间,同时也利于硬件实现。

Research on the algorithms of iris recognition based on single frequency 2D-Gabor filter

For the problem of the 2D-Gabor filter design, a method of single frequency 2D-Gabor filter design is proposed. The meaning of every parameter of 2D-Gabor filter in detail is analyzed, and the single frequency ω fit for the iris texture is found through dichotomy, and finally every parameter is chosen according to the relationship between ω. One 2D-Gabor filter in single frequency is designed. Experimental results show that when the proposed designed single frequency 2D-Gabor filter is used to perform feature extraction, the recognition rate could be above 98.61%. So taking advantage of single frequency 2D-Gabor filter not only can increase the speed of feature extracting, decrease memories but also can be convenient for hardware to realize.