Iris recognition using Level Set and hGEFE
Joseph Shelton, Kaushik Roy, Foysal Ahmad, Brian O'Connor · 2014
In this paper, we deploy a Fuzzy C-Means Clustering with a Level Set (FCMLS) method in an effort to localize the nonideal iris images accurately. We apply Genetic and Evolutionary Feature Extraction (GEFE), a method that evolves Local Binary Pattern (LBP) based feature extractors in order to elicit the most discriminating biometric features. In addition, a hybrid Genetic and Evolutionary Feature Weighting/Selection (GEFeWS) method is applied to select and weight the most important features. GEFeWS uses a genetic and evolutionary computation (GEC) to evolve a population of real-coded feature masks (FMs). We apply GEFeWS on features extracted by GEFE, and we refer to this technique as hybrid GEFE/GEFeWS, or hGEFE. Results show that hGEFE provides a significant increase in recognition accuracy while reducing the number of features being used when compared to just using GEFE alone.