基于离散Curvelet变换和LS-SVM的虹膜特征提取与识别

提出了一种基于离散曲波变换和最小二乘支持向量机(LS-SVM)的虹膜特征提取与分类识别的新方法。对虹膜纹理采用离散Curvelet变换,提取低频子带系数矩阵的均值方差和高频子带能量作为虹膜图像的特征向量,利用最优二叉树多类LS-SVM分类器进行分类与识别。MATLAB仿真实验结果表明,与现有方法相比,该算法识别准确率较高,能有效应用于身份认证系统中。

Feature extraction and recognition of iris based on discrete Curvelet transform and LS-SVM

A novel method for iris feature extraction and recognition is proposed by integrating dis-crete Curvelet transform and least square support vector machine( LS-SVM) . An iris image is convolved by discrete Curvelet transform. Extracted mean and variance of low frequency sub-band coefficients and the energy of high frequency sub-band are used to represent feature vectors of iris image; LS-SVM with optimal binary tree is developed to implement classification and recognition. The experimental results of the simulation with MATLAB show that the proposed algorithm has higher iris recognition accuracy rate than present method and can be used for a personal identification system in an efficient manner.