Performance analysis on iris recognition based on half polar iris localization and normalization method using modified low cost camera

摘要

Iris recognition offers unique pattern as a biometric authentication. Iris has advantages in term of universality, distinctiveness, permanence, and collectability. This research proposed implementation of half polar iris localization and normalization to improve performance of iris recognition that is detected using modified low cost camera. In the development phase dataset from CASIA-IrisV1 is used. The performance of the recognition is then evaluated based on level of accuracy and execution time. Our proposed method produces better accuracy for Iris Localization, and hence suitable for Iris Recognition The processing speed analysis shows that the proposed method performs faster Iris based authentication.