Find all the Top AICybersecurity papers. Links to pdf, code repos and demos are provided.
Presents ML-LOO, a novel framework for detecting adversarial examples in deep neural networks by leveraging feature attribution methods. It demonstrat...
see more
Introduces HopSkipJumpAttack, a novel framework for decision-based adversarial attacks that seeks to create adversarial examples by leveraging output ...
see more
Contactless palmprints are comprised of both global and local discriminative features. Most prior work focuses on extracting global features or local ...
see more
Evaluating the risk level of adversarial images is essential for safely deploying face authentication models in the real world. Popular approaches for...
see more
Spoof detectors are classifiers that are trained to distinguish spoof fingerprints from bonafide ones. However, state of the art spoof detectors do no...
see more
Face image quality can be defined as a measure of the utility of a face image
to automatic face recognition. In this work, we propose (and compare) tw...
see more
Introduces AdvFaces, a novel adversarial face synthesis method that leverages Generative Adversarial Networks (GAN) to create imperceptible perturbati...
see more
Presents an advanced method for detecting fingerprint spoof attacks using deep convolutional neural networks (CNNs). The proposed approach achieves st...
see more
Explores the critical challenges faced by biometric recognition systems, which, despite their advantages, have raised significant concerns regarding t...
see more
Optical coherent tomography (OCT) fingerprint technology provides rich depth information, including internal fingerprint (papillary junction) and swea...
see more
Prevailing evaluations of fingerprint recognition systems have been performed as end-to-end black-box tests of fingerprint identification or authentic...
see more
Prevailing user authentication schemes on smartphones rely on explicit user
interaction, where a user types in a passcode or presents a biometric cue ...
see more
