Fusion of Iris and Periocular Biometrics Authentication using CNN
Nalluri Prasanth, Chilukuri Uday Sai Kiran, Sahithi Nethi, Trishitha Ketineni, N. Srinivasu, G. Pradeepini · 2023
Iris recognition, which is one of the most accurate biometrics, plays a very important role in authentication systems. Many iris recognition methods are inaccurate since the iris region is unclear for authentication. This paper focuses on the eye region along with the periocular region for biometrics. Periocular biometrics includes eye regions like the face which surrounds the eyes, including the eyelids, eyelashes, and eyebrows. The periocular region has been found to offer greater recognition rates in challenging image-capturing conditions compared to the performance of iris recognition algorithms due to poor iris segmentation, image blur, contours, and partial occlusion from eyelids and eyelashes. This paper is based on deep learning, using CNN for Iris Recognition. Based on the publicly available datasets CASIA-Iris-Thousand and UBIRIS.v2 databases, this proposed method has shown significant improvement in accuracy.