Research on iris image preprocessing algorithm
Pengfei Zhang, Qiuming Li · 2005
A fast algorithm for iris image preprocessing was proposed in the paper. The based fuzzy linear Hough transform eyelid detection method is presented. By pupil biology character, gradient pair vector (GPV) based on center symmetry rule was used to search the inner circle and obtain the circle parameters through the statistic calculating. The outer circle detection made use of the proposed RCD method without accumulator. And the RCD is faster than the Hough circle detection method in the no-heavy noise. The integro-difference operator was employed to accomplish the coarse to exact iris localization. The iris normalization was invariant for translation, rotation and scale after mapping into polar coordinates. Experimental and results show that the proposed method has an encouraging performance, shows 98.7% localization success and reduces the system operation time. The preprocessing algorithm is robust in noise.