Rotation Invariant Feature Extraction and Matching Methodology for IRIS Recognition
Monika, Maroti Deshmukh · 2018
In this multimedia era, the iris data are collected and recorded by the use of hardware devices like camera for biometric authentication system. This produces a huge amount of iris image data which requires an efficient authentication system. A significant work has been done for authentication systems, although the performance of them is far behind to fulfill the stated requirement. In this paper, we propose an effective and efficient iris recognition system for individual verification and identification. The proposed system utilizes a novel rotation invariant feature extraction technique with a sophisticated feature-matching methodology to solve the problem in robust environment. The experimental results of the proposed system shows its effectiveness and robustness in different standard iris-databases. Three widely well known databases (IIT Delhi database, MMU1, MMU2) are used for the experimental purpose. The mean-precision, mean-recall, mean-F-measure, and accuracy are used as the performance measure in this work.