The Eye's Signature: Innovative Approaches to Iris Detection

Abstract

This research aims to develop and evaluate a deep learning-based iris detection system using a specialized Convolutional Neural Network (CNN) architecture. The research methodology includes data set preprocessing, CNN model design, training using Adam optimization, as well as evaluation using accuracy, precision, recall, and F1 score metrics. The dataset used was obtained from Kaggle and preprocessed before being divided into training, validation, and testing sets. The CNN model consists of three convolutional layers with increasing filter sizes (32, 64, and 128), ReLU activation, batch normalization, and MaxPooling layers for efficient feature extraction, as well as dropout regularization to reduce overfitting. Experimental results show that the proposed model achieves a high classification accuracy of 97.33%, with robust performance against variations and noise in iris images. Comparative analysis with traditional iris recognition methods confirms the superiority of deep learning in handling challenges such as lighting changes and occlusions. Although the results are promising, challenges such as data bias and computational demands are still a concern. Future research will explore more advanced architectures as well as additional pre-processing techniques to improve the generalizability and effectiveness of the system in real-world applications.