Abstract:In the modern era of computing, the iris-based biometric systems are gaining significant attention for secured and automatic human authentication. However, past decades have witnessed numerous spoofing assaults on these iris-based recognition systems where an attacker impersonates an exact replica of biometrical information of the genuine user. Particularly, these direct attacks are targeted on the iris sensor module of the biometric system by presenting the fake artefacts of a bonafide iris trait. With the emergence of data-driven paradigm (i.e. handcrafted feature learners such as support vector machine (SVM), decision tree (DT), k-nearest neighbor (KNN), ensembles, etc. or automatic image features extraction-based classifiers such as convolutional neural networks (CNN), generative adversial networks (GAN)), mitigating these iris spoof attacks has become comparatively an easier and accurate task of computer vision. An iris spoof detector (ISD) is a mechanism through which the vitality of a presented iris trait is measured intelligently by classifying it as genuine or counterfeit. In this study, we explicate a taxonomy-based comparative analysis of state-of-the-art (SOTA) ISDs that employ machine learning or deep learning-based approaches. We expound a novel taxonomy for classifying ISDs based on underlying criterion such as feature type, learning algorithm, pre-trained models, data augmentation, hybrid, etc. Furthermore, we investigate and analyze various benchmark datasets employed in the various data-driven iris spoof detectors (D2ISD). We also illustrate prominent performance evaluation protocols that are widely adopted in the SOTA approaches. Though, pioneer contributions related to D2ISD is reported in the literature, but several potential open research problems still exist, that requisite a futuristic attention of the investigators in this active field of research.

A Comprehensive survey on Iris Presentation attacks and Detection based on Generative Adversarial Network

Introduction to Presentation Attack Detection in Iris Biometrics and Recent Advances

Synthesizing Iris Images using Generative Adversarial Networks: Survey and Comparative Analysis

Iris Presentation Attack: Assessing the Impact of Combining Vanadium Dioxide Films with Artificial Eyes

A survey on data-driven iris spoof detectors: state-of-the-art, open issues and future perspectives

Iris Liveness Detection: A Survey

Generalized Iris Presentation Attack Detection Algorithm under Cross-Database Settings

A Parametric Approach to Adversarial Augmentation for Cross-Domain Iris Presentation Attack Detection

Presentation Attack Detection: A Systematic Literature Review

Presentation Attack Detection using Convolutional Neural Networks and Local Binary Patterns

Deep convolutional neural networks for face and iris presentation attack detection: Survey and case study

Presentation Attack detection using Wavelet Transform and Deep Residual Neural Net

Iris presentation attack detection based on best-k feature selection from YOLO inspired RoI

Viability of Optical Coherence Tomography for Iris Presentation Attack Detection

IRIS TECHNOLOGY: A REVIEW ON IRIS BASED BIOMETRIC SYSTEMS FOR UNIQUE HUMAN IDENTIFICATION

Privacy-Safe Iris Presentation Attack Detection

Multimodal Security of Iris and Fingerprint with Bloom Filters

Open Source Presentation Attack Detection Baseline for Iris Recognition

A Novel Active Solution for Two-Dimensional Face Presentation Attack Detection

Introduction to Presentation Attack Detection in Face Biometrics and Recent Advances

Alpha-wolves and Alpha-mammals: Exploring Dictionary Attacks on Iris Recognition Systems