Survey on Neural Networks for Presentation Attack Detection (NNP)

EasyChair Preprint 10114

Abstract

In many different applications, including security, access control, and identity verification,
facial recognition systems are becoming more and more common. These systems are, however,
susceptible to presentation assaults, in which a perpetrator tries to get past the defenses by presenting
a fictitious or changed image of a face. Different Presentation Attack Detection (PAD) techniques
have been created to identify presentation attacks in facial recognition systems in order to solve this
issue. Since they can learn distinguishing features and model intricate relationships in the data, neural
networks have become a promising PAD method for facial recognition in recent years. The classic
methods for PAD in facial recognition as well as the various presentation assaults are described in this
comprehensive work.

Keyphrases: Convolutional Neural Networks (CNNs), Generative Adversarial Networks (GANs), Presentation Attack Detection (PAD), Recurrent Neural Networks (RNNs), Security, facial recognition systems, neural networks

@booklet{EasyChair:10114,
  author    = {Willy Kinfoussia},
  title     = {Survey on Neural Networks for Presentation Attack Detection (NNP)},
  howpublished = {EasyChair Preprint 10114},
  year      = {EasyChair, 2023}}