• K A S S Mounika
• Dr. A M Prasad

Iris recognition is one of the key processes in verifying a person’s identity. To get the accurate information about the identity of the person through it, iris segmentation plays a major role. With the increase in deep learning, researches on automatic iris segmentation schemes are increasing day-by-day. In this paper, an automatic iris segmentation scheme is proposed using the SEGNET model. This model is trained and tested on CASIA-Iris-Interval v4.0 dataset from [10] and segmentation results are obtained. The accuracy and errors are obtained by calculating 3 different parameters namely F1 score, Nice1 and Nice2 scores respectively. By using this method, it is observed that this model gives the results with better accuracy of 99.6% and the minimum Nice1 and Nice2 scores such as 0.6% and 0.3% respectively. As this model consists of number of training parameters, it is considered to be better and feasible model when compared to other Convolutional Neural Network (CNN) models such as U-Net.

Iris Recognition, Iris Segmentation, Deep Learning, Convolutional Neural Networks, U-Net, SEGNET

"A Robust Iris Segmentation Method Using Deep Neural Networks", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.9, Issue 9, page no.e693-e697, September-2022, Available :http://www.jetir.org/papers/JETIR2209482.pdf

"A Robust Iris Segmentation Method Using Deep Neural Networks", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.9, Issue 9, page no. ppe693-e697, September-2022, Available at : http://www.jetir.org/papers/JETIR2209482.pdf