Advanced Biometric Iris Authentication System Leveraging Deep Learning Techniques

This paper provides an advanced biometric iris authentication system using deep learning, particularly Convolutional Neural Networks (CNNs), in order to enhance the performance as well as efficiency of iris-based biometrics. The proposed system overcomes a few of the shortcomings of traditional methods of iris authentication, such as susceptibility to spoofing and reduced accuracy when lighting conditions change. A very large dataset of iris images was preprocessed to minimize variations, ensuring the best possible accuracy in recognition of iris patterns. The system outperformed the state-of-the-art systems in terms of False Acceptance Rate (FAR) and False Rejection Rate (FRR). Furthermore, high-level spoofing detection capabilities were added, significantly lowering vulnerability to attacks using fake iris images. One of the significant contributions of this work is creating a particular CNN architecture for iris recognition and a novel preprocessing pipeline for more improved feature extraction. Results show that the proposed deep learning-based approach outperforms the already existing systems, so it is a promising solution for secure and privacy-preserving human identification in real-world applications. This research will serve as a seed to future studies in fine-tuning deep learning models applied to different security domains in biometric systems.