Designing a New Adaptive Filter for Iris Feature Extraction

Abstract

With a growing emphasis on human identification, iris recognition as a biometric identification has recently received increasing attention. In these systems, matching error is of utmost importance. In this paper for feature extraction of iris a new adaptive filter is designed. This filter represents the texture features of iris more accurately. The scope and directions of this filter changes depending on situation and texture variations around each pixel. Finally extracted feature by some methods of texture analysis are combined. In order to reduce the lengths of feature vectors, PCA is imposed locally on them. For matching operation a new approach is proposed as well. The simulation results represent the better performance of our new designed system with reduced matching error rate tending to zero.