Abstract

Human iris’ identification is a constantly developing technology and it has it’s own significant in many commonplace applications such as financial sector, identity verification, evidence analysis, law enforcement, and security standards. Several obstacles face the recognition of the iris and the high variation in its captured image is one the most highly affected that is brought on by many factors including aging, illumination, and occlusion. Furthermore, there are some issues with the computing time and complexity of systems concerned in recognizing iris that require attention. In this research, a proposed Iris recognition system that can show a high recognition accuracy and a reduced time is presented. The Corona Virus Optimization Algorithm is a sophisticated bioinspired algorithm that serves as the foundation for the suggested system. The main objective of the suggested approach is to increase the iris identification accuracy rate by fi-ne-tuning the hyperparameter of six conventional Machine Learning models and selecting as well refining the most useful features. Four versions of Iris Image Database known as of CASIA (i.e., 1.0, 2.0, 3.0, 4.0), have been employed to test the system. The evaluation experiment outcomes findings proven the system’s efficiency in catching the high recognition accuracy in uncontrolled environments when compared to current methods. This is accomplished in a through a recognition time ranging from 1564.16 to 13.97 milliseconds, requiring extraordinarily little processing complexity and effort to attain 94%–100% accuracy. 

Statistics