Method for generating Cancelable iris template by utilizing binary filter

Traditional human authentication systems have passive issues which were overthrown by biometric-based recognition systems. Still, templates in biometric recognition systems undergo passive issues like security and privacy invasion. To avert the breach of biometric patterns, there is need to alter the actual biometric pattern before accumulating in the database through irreversible or non-invertible operation. The altered template is termed as Cancelable template which perhaps be reissued or replaced in the event of compromise. Although various cancelable biometric schemes were suggested but with extensively diminishing accuracy. In this paper, we come up with novel cancelable iris technique stationed on arbitrary binary filter. Feature vector of iris is formed by utilizing 1-D Log Gabor Filter which is enforced on iris image. Pre-alignment is done to overwhelm rotational invariance and row vector is created from it. A decimal vector is generated after xoring a random vector with the row vector. Cancelable iris template is formed by mapping the decimal values to corresponding indices of binary filter. To affirm the accuracy of our suggested approach, experimentation is done on distinct iris datasets. After employing the approach, we attained Equal Error Rate (EER) of 0.98% for CASIA-V 1.0 and 2.6% for CASIA-V3-Interval iris datasets respectively. Furthermore, the technique sustain the revocability, unlinkability and unlinkability attributes of cancelable template.