2020/07/02 by Renu Sharma, Arun Ross, Sharma, Renu +1 · 2 citations
Computer Science · #Biometric Identification and Security #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences
paper · pdf · doi:10.48550/arxiv.2007.01381
openalex publication_date 2020/07/02 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
An iris recognition system is vulnerable to presentation attacks, or PAs,\nwhere an adversary presents artifacts such as printed eyes, plastic eyes, or\ncosmetic contact lenses to circumvent the system. In this work, we propose an\neffective and robust iris PA detector called D-NetPAD based on the DenseNet\nconvolutional neural network architecture. It demonstrates generalizability\nacross PA artifacts, sensors and datasets. Experiments conducted on a\nproprietary dataset and a publicly available dataset (LivDet-2017) substantiate\nthe effectiveness of the proposed method for iris PA detection. The proposed\nmethod results in a true detection rate of 98.58 % at a false detection rate of\n0.2 % on the proprietary dataset and outperfoms state-of-the-art methods on the\nLivDet-2017 dataset. We visualize intermediate feature distributions and\nfixation heatmaps using t-SNE plots and Grad-CAM, respectively, in order to\nexplain the performance of D-NetPAD. Further, we conduct a frequency analysis\nto explain the nature of features being extracted by the network. The source\ncode and trained model are available at https://github.com/iPRoBe-lab/D-NetPAD.\n