Kernel Clusterand SVMs-Based Algorithm for Iris Rough Classification in Massive Databases

The kernel method was employed into an index algorithm on iris recognition, while clustering the massive databases under unsupervised learning. And this algorithm was certified to have a good performance in iris classification from large-scale databases by Support Vector Machines. First of all, we proposed three criterions of coding iris images in application to index. According to these requirements, we presented an algorithm on extracting statistical features from wavelet coefficients. Before matching iris codes, we cluster the iris databases by unsupervised learning based on kernel methods. In the end, the clustering algorithm was verified by using SVMs in CASIA 2.0 and a set of synthetic data. Experimental results show that the clustering method we proposed has a better performance and shortens the runtime of the system.