Improving Iris Recognition Performance using Local Binary Pattern and Combined RBFNN
Kamal Hajari
Kamal Hajari, Department of Computer Technology, Yeshwantrao Chavan College of Engineering, Nagpur (Maharashtra), India.
Manuscript received on 15 April 2015 | Revised Manuscript received on 25 April 2015 | Manuscript Published on 30 April 2015 | PP: 108-112 | Volume-4 Issue-4, April 2015 | Retrieval Number: D3903044415/15©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Biometric is constantly evolving technology due to increased concerns in security. It exploits discriminable behavioral or physiological characteristics to identify a legitimate individual. The physiological features like DNA, Iris, Retina, Palm print, face, Ear, Fingerprint and Hand geometry etc. are being extensively used as biometric features to discriminate among different individuals. Iris recognition is a challenging problem, because iris is distinct and intrinsic organ, which is externally visible and yet secured one. It is well protected by the eyelid and the cornea from environmental damage. Our primary focus is to develop reliable system and increase the iris recognition rate on CASIA iris dataset. In this paper, a novel texture features are derived from iris images using histogram of Local Binary Pattern (LBP) and the Neural Network based classifier, namely Radial basis function networks is implemented for classification. Before feature extraction, pre-processing of iris images is performed including iris localization, Segmentation and Normalization. The proposed system give high recognition rate of 93.5% on CASIA iris dataset compared with other methods.
Keywords: Local Binary Pattern, Radial basis Neural Network Classifier, CASIA, Histograms
Scope of the Article: Pattern Recognition