Improved Iris Biometric Identification Through Hybrid Stochastic Search and Shark Smell Optimization

The retina is regarded as the most reliable and precise biometric because of its uniqueness and unchangeability. It is a critical factor that requires high accuracy and speed during recognition. Therefore, new iris recognition methods reduce the region of interest (ROI) size while maintaining high accuracy. This paper proposes an iris recognition model for biometric authentication. Shark-smell optimization (SSO) is used in this model to extract retina features as a biometric. SSO specifically targets the ROI of the retina. The ROI is identified using the Hough transform, which requires consistent features and input parameters to control the results. Additionally, the Convex Hull is used to determine the area around the SSO points. The model's accuracy is evaluated using a probabilistic neural network (PNN) to verify the consistency of the features extracted. The PNN achieves a classification accuracy of 99.74% on the Casia V2.0 dataset. The experimental results demonstrate that the features of each input image remain consistent when processed with the same parameters.