Pengenalan Iris Menggunakan K – Nearest Neighbors dengan Ekstraksi Fitur Dicrete Cosine Transform

Iris is a biometric based on physiological characteristics which are regarded as highly reliable in biometric recognition systems. The iris pattern between one person and another is very different, identical twins have different iris patterns, so the recognition system using iris has a very good level of security. In this research proposed iris recognition system using K-Nearest Neighbors as classifier and Discrete Cosine Transforms as feature extraction algorithm. The noisy regions should be distinguished before feature extraction in a pre-processing stage called segmentation (Localization and noise-removing) and normalization. The normalization is a transform from Cartesian to polar coordinates. The iris image data used as a training image and test image are public datasets with a total data of 420 iris images. The experiment results show the level of recognition accuracy is 70%