Iris texture feature extraction with orthogonal polynomials

In this paper, a feature extraction technique with orthogonal polynomials based computational model to accurately extract local texture in iris images is presented. Initially, the normalized input iris image is subjected with the orthogonal polynomials model and the model coefficients are obtained. The model coefficients are subjected to statistical hypothesis testing with Hartley’s test so as to extract the signal components due to texture in the iris images and simultaneously separating out the noise components. These model coefficients due to the orthogonal polynomials model, are utilized to represent the iris texture patterns along with their zonal positions, as the locations of the micro texture present in the image analysis is considered to be significant. The texture primitives thus extracted are represented with a decimal number and used for feature extraction.