Iris Recognition using Orthogonal Transforms

Transforms M.Mani Roja , Dr.Sudhir Sawarkar , # Research scholar, Amravati University Associate Professor, TSEC, Mumbai, India. 1 [email protected] * Principal, DMCE Navi Mumbai, India 2 [email protected] Abstract— Iris Recognition is a biometric recognition technique in which features of the iris are used to uniquely identify individuals. Iris recognition has over the years emerged as one of the most accurate biometric techniques as opposed to other biometric techniques like face, signature and fingerprint. First, the iris image is pre processed using canny edge detector using a Gaussian filter. The iris edge and the pupil edge are extracted using image morphological operation, image opening. After normalization of red, green and blue components of the colour iris using Euclidean distance method, they are combined to form the localized colour iris. For feature vectors extraction, orthogonal transforms like discrete cosine transform, discrete sine transform and discrete Fourier transform have been considered. The proposed iris recognition system is very time efficient and it takes less than 1 second to grant authentication. KeywordImage processing, Discrete Cosine transform Transform, Discrete Sine Transform, Discrete Fourier Transform.