基于改进的Log-Gabor小波的虹膜识别算法

对二维Gabor复小波应用于虹膜识别进行了详细的理论分析,提出了采用二维奇对称Gabor小波代替二维Gabor复小波来提取虹膜纹理特征的改进算法.在此基础上,进一步提出一种采用改进的二维 Log-Gabor小波来提取虹膜纹理特征,采用汉明距来进行特征匹配的新方法,克服了二维Gabor复小波应用于虹膜识别的缺陷.与已有算法进行比较的实验数据表明,采用二维奇对称Gabor小波的改进算法在识别率略有提高的基础上,能大大地减少编码存储空间以及编码和匹配的时间.而采用改进的二维Log-Gabor小波算法则进一步的提高了识别率.

Iris Recognition Using Modified Log-Gabor Wavelets

In this paper, based on a thorough theoretical iris recognition, an improved method, which uses the 2D analysis of the complex 2D Gabor wavelets for odd Gabor wavelets to replace the original 2D Gabor wavelets, is first proposed for iris texture extraction. After that, a new method for iris recognition is further proposed by using a modified 2D Log-Gabor wavelet for iris texture extraction, followed by a Hamming Distance based matching to overcome the shortcomings of the complex 2D Gabor wavelets. Experimental results show that the method based on the 2D odd Gabor wavelets can substantially reduce the storage load and increase the matching efficiency besides a slight increase of iris recognition rate in comparison with the complex 2D Gabor wavelets based method. In addition, the modified 2D Log-Gabor wavelets based method is shown to be able to further increase the iris recognition rate markedly.