Iris based biometric identification system

Widespread internet usage has led to data protection and accurate verification of personnel, resulting in extensive use of biometrics. In the proposed work, we have discussed Daugman´s algorithm, a commercially prevalent iris biometric system. The main aim is to compute false accept rate (FAR) and false reject rate (FRR) for any standard iris database. Initial stages involve removal of specular reflections, thresholding & elimination of local minima. After segmenting iris using Daugman´s operator, it is normalized into a dimensionally constant rectangular block. Feature extraction is done using Gabor filter. Hamming distance is employed for matching binary encoded iris templates, and two templates are a match if the mismatch ratio is less than 0.27. All the simulations have been performed using MATLAB®. The algorithm was applied for 995 images of MMU2 iris database. The accuracy of the algorithm for segmentation is approx. 86.94% i.e. 865 out of 995 segmented successfully. For 3 iris samples, FAR is 3.93% and FRR is 0.12%.