Enhancement of Accuracy for Iris Presentation Attack

Biometrics is the assessment and statistical examination of an individual's distinctive physical and behavioural traits. Identification, access control, and identifying those who are being monitored are the major areas where technology is focused. In security systems, biometric-based recognition has been substituting traditional recognition techniques. Iris recognition (IR) is more utilized in today's biometric technologies, which are applied to various devices for various security concerns. The effectiveness of IR systems has significantly improved during the past ten years. This is due to recent developments in deep convolutional neural networks (CNNs) and computer vision, as well as the availability of enormous training datasets. A presentation attack is a kind of attack where an imposter creates artificial or fake out of genuine. This paper proposed an effective approach to increase the accuracy in detecting iris presentation attacks and design the state-of-the-art of existing CNN techniques from 2015 to 2023. The proposed method is a Dual Channel Convolutional Neural Network Presentation Attack Detector(DC-CNNPAD) for increasing the accuracy of detecting real iris. The experiment was performed on the sample LivDet-2017 dataset to identify the effectiveness of the model in order to detect the artifact. The results obtained from the detection model on the sample dataset are extremely good.