Iris Recognition Using Principal Component Analysis

摘要

Iris recognition is used in several fields such as security because its singularity and stability. In this research, Principal Component Analysis (PCA), is adopted and applied to the CASIA V3i database images after some preprocessing technique like starching. The results shows that the best matching appears when 40 column and five components are selected with 6.018 Euclidean distance (93.982%). To find Iris Sheet, new method is suggested and named smooth median filter as preprocessing method then the result is normalized Dougman rubber sheet to convert from Cartesian to polar coordinate to find the iris/pupil boundary and sclera/iris boundary.