Level set approaches and adaptive asymmetrical SVMs applied for nonideal iris recognition

Abstract

In this paper, we present algorithms for iris segmentation, feature extraction and selection, and iris pattern matching. To segment the nonideal iris images accurately, we propose level set based curve evolution approaches using the edge-stopping function and the energy minimization algorithm. Daubechies Wavelet Transform (DBWT) is used to extract the textural features, and Support Vector Machine-Recursive Feature Elimination (SVM-RFE) algorithm is deployed to reduce the feature dimension without compromising the accuracy. To speed up the matching process and to control the misclassification error, we apply a combined approach called Adaptive Asymmetrical SVMs (AASVMs). The verification performance of the proposed scheme is validated using the UBIRIS Version 2, the ICE 2005, and the WVU datasets. © 2009 Springer Berlin Heidelberg.