IRIS feature vector reduction using haar wavelet transform and local binary pattern

Abstract

Iris recognition is a means of biometric identification. A key part of the recognition system using iris is the extraction of prominent texture information or features in the iris. The identification delay in iris recognition can be reduced by reducing the feature vector generated from the feature extraction of iris images. In this thesis, two algorithms are proposed for the reduction of the iris feature vector. The first method is the combination of Haar wavelet transformation (HWT) and local binary pattern (LBP) termed here as HWT-LBP. The second method is a new form of LBP termed LBPX. First, HWT-LBP is considered. In this case, input eye images are processed and converted to normalized iris images employing circular Hough transformation and Daugman’s rubber sheet model. HWT is then applied to the normalized image. The output of this HWT goes through the LBP process. In this hybrid method, HWT is applied to the normalized iris image resulting in four output images, including the approximation image known as LL sub-band. This LL sub-band is then further decomposed using HWT into four sub-images. The resultant second-level LL is decomposed using HWT into the third-level LL sub-band. The application of repeated HWT extracts the major information containing region, reducing the information size. Next, MLBP is applied to the obtained LL, where MLBP includes LBP and XOR operations. The output of MLBP is a binary iris template. The effectiveness of this proposed hybrid HWT-MLBP method is experimentally evaluated using three different datasets, namely CASIA-IRIS-V4, CASIA-IRIS-V1 and MMU. The proposed HWT-MLBP method can obtain a reduced feature vector length of 1×64. For instance, when applied to CASIA-IRIS-V1 dataset, HWT-MLBP can obtain an average correct recognition rate of 98.30% and false acceptance rate of 0.003%. Results indicate that the proposed HWT-MLBP outperforms existing methods in terms of reduced feature length, which ensures faster iris recognition. Next, the LBPX method is considered. The LBPX method is based on the concepts of uniform LBP, rotation-invariant LBP, and XOR operators. Moreover, the existing rotation-invariant uniform LBP (RIU LBP) method is applied here in the context of iris feature extraction. LBPX is applied to the normalized iris images. The performance of LBPX based recognition system adopting iris image is evaluated in terms of accuracy and feature vector length. This is done for three datasets CASIA-IRIS-V4, UBIRIS and IITD. Results indicate that LBPX can achieve acceptable accuracy values of 97.15%, 97.20%, 96% and 96.40% for CASIA-IRIS-V1, CASIA-IRIS-V4, UBIRIS, and IITD datasets, respectively. Furthermore, results show LBPX outperforms existing feature extraction methods in terms of reduced feature-length, ensuring faster iris recognition.