Feature Extraction for Iris Recognition

In this paper, we evaluate the performance of feature extraction methods for iris pattern classification. Generally, the identification system using iris recognition consists of the iris localization block and the iris pattern classification block. In this paper, we used the 2D bisection-based Hough transform and the radius histogram method for the iris localization and we used multilayer perceptrons for the iris pattern classification. Then, the three linear feature extraction methods are evaluated to reduce the classification time and system complexity. The evaluated feature extraction methods are the feature extraction based on decision boundary, the canonical analysis, and the principal component analysis. Experiments with 1831 iris images show that the feature extraction based on decision boundary and the canonical analysis show a favorable performance for the iris recognition.