ContactLensIris: A Robust Iris Recognition System based on Integrated Deep Learning Approach

The popularity of iris-based biometric systems used for authentication systems is increasing day by day and some covariates such as contact lenses, aging, etc. reduce their efficiency. So, this work aims to develop an efficient recognition system that ensures the performance of the proposed system in the presence of contact lenses. This proposed system named ContactLensIris, first enhances image quality using median filter and contrast enhancement. Segmentation is used to separate the iris and pupil and this work uses histogram equalization. Features of the segmented iris image are extracted using ORB, BRISK, and CNN, fuse all these features followed by a fully connected layer, and compute the classification probability for final results. The proposed system is first trained with 70% of the data of NDCLD15 and then the rest 30% is used for testing and validation. The obtained results of the proposed ContactLensIris ensure its effectiveness with 98% CCR and 1.6% EER.