ROTATION COMPENSATION IRIS RECOGNITION SYSTEM USING MULTI-SCALE DIRECTIONAL FILTER BANK AND 2DPCA

Chellakkon Helen Sulochana and Srinivasan Selvan

Keywords

Biometrics, iris recognition, matching, multiscale directional filter bank, 2DPCA

Abstract

This paper proposes a novel rotation invariant iris recognition method using the eigen features of directional filter bank (DFB). Rotation invariant feature vectors are obtained by applying Eigenanalysis to the directional sub-bands. Experimental results show that DFB captures the directional information of iris images and the rotation invariant features achieve high classification accuracy on rotated images (due to head tilt) of the database.

Important Links:



Go Back