Advancing Security on Harnessing Deep Learning for Iris Recognition and Classification
K. Sivasankari, D. Kerana Hanirex · 2025
This chapter thoroughly examined IR systems using DL methodologies to augment security and authentication protocols. A range of deep learning methods, including CNNs like ResNet and DenseNet, were used to enhance the accuracy of iris segmentation and identification. The study's results demonstrate notable progress in the accuracy of IR detection, as shown by an average identification rate of 98.5% across several datasets. The integration of IR technology with other biometric modalities, including fingerprint identification, has shown enhanced security measures, resulting in an integrated accuracy rate of 99.15%. Furthermore, the study highlights the capacity of DL-based IR systems to overcome conventional obstacles, such as fluctuations in lighting and obstructions, resulting in strong and dependable authentication solutions. The significance of using DL algorithms in security applications is underscored by these improvements, which provide improved cyber hygiene and access control mechanisms.