Iris Recognition System Using Combined Colour Statistics

This paper proposes a high performance iris recognition system based on the probability distribution functions (PDF) of pixels in different colour channels. The PDFs of the segmented iris images are used as statistical feature vectors for the recognition of irises by minimizing the Kullback-Leibler distance (KLD) between the PDF of a given iris and the PDFs of irises in the training set. Feature vector fusion (FVF) and majority voting (MV) methods have been employed to combine feature vectors obtained from different colour channels in YCbCr and RGB colour spaces to improve the recognition performance. The system has been tested on the segmented iris images from the UPOL iris database. The proposed system gives a 98.44% recognition rate on that iris database.