DEVELOPMENT OF A HUMAN IDENTITY VERIFICATION SYSTEM USING MULTIMODAL BIOMETRICS

Abstract

Identification systems based on something other than an intrinsic aspect of a human being are not always secure; thus, biometrics system deployment. Conversely, most biometrics systems deployed in the real world are unimodal. They rely on a single biometric characteristic which could be hacked or spoofed by modern burglars. This study presents a Coded Wavelet Processing (CWP) fusion algorithm for Iris, Nose and Ear multimodal biometrics verification. The design processes each Iris, Ear and Nose dataset through a common pipeline of multi-algorithms of preprocessing, Feature extraction and analysis. At the preprocessing phase, dimension reduction, geometric transformation and illumination adjustment of the images were done. Features were extracted using Haar wavelet decomposition of the image at the second level (low-low) decomposed sub-band. The resulting image was further encoded using two algorithms namely: Fast Walsh-Hadamard Transform (FWHT) and Singular Value Decomposition (SVD). The overall extracted features of Iris, Nose and Ear were fused using concatenation technique; the result is stored as one dimension feature vector; recognition rate using appropriate threshold for each modality was 91%. Furthermore, in order to carry out the performance evaluation of the CWP based system, a benchmark algorithm based on Hough Transform, Principal Component Analysis and Wavelet Decomposition (HPW) for Iris, ear and nose respectively were designed; owing to the fact that, there was no previous work based on CWP algorithm. The design was simulated using the same dataset, a recognition rate of 72% was observed with reduced processing speed when compared with that of CWP. In view of the above, the developed recognition system using Iris modality will be affordable even to nations with low or medium income economy.