Introduction to Presentation Attack Detection in Face Biometrics and Recent Advances

Hernandez-Ortega, Javier; Fiérrez, Julián; Morales, Aythami; Galbally, Javier

doi:10.48550/arxiv.2111.11794

Cited by 4 publications

(3 citation statements)

References 66 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The discovery is essential as the latest models with the most exact results depend on deep-fake learning techniques comprising numerous neural organization boundaries. Reasonable AI in PC vision in this way is an examination heading that is expected to advance and use the advances and focal points of AI and AI in computerized media legal sciences [37][38][39][40][41][42][43][44][45].…”

Section: The Discussionmentioning

confidence: 99%

A robust facemask forgery detection system in video

Almukhtar¹

2022

PEN

View full text Add to dashboard Cite

An in-depth fake video uses an Artificial Intelligent (AI), AI programming, and a Personal computer (PC) mix to create a deep fake video of the action. Deep-faking can also be used to represent images and sounds. We provide insights into our reviews in this document. We're showing our dataset to start. At this point, we present the subtleties and reproductively of exploratory settings to evaluate the discovered effects finally. It is no surprise to find deep fake videos, which only monitor a tiny section of the video (e.g., the target face appears quickly on the video; hence the time is limited). We remove our system's fixed duration's persistent effects as each video contributes to the preparation, approval, and testing sections to reflect this. The edge groups are isolated from each video successively (without outline skips). The entire pipeline is ready to be finished when the approval stage is ten years old. Convolutional Neural Network (CNN) was the best and most reliable of the classification systems. Fake videos typically use low-quality pictures to mask faults or insist that the general public regard camera defects as unexplainable phenomena. 'This is a common trope with Unidentified Flying Object (UFO) videos: ghostly orbs are lenses; snakes are compression artifacts on one's face. In this study, we have implemented a sophisticated, knowledgeable method to recognize false images. Our test results using various monitored videos have shown that we can reliably predict whether videos are monitored through with simple co-evolutionary Long Short-Term Memory (LSTM) structure.

show abstract

Section: The Discussionmentioning

confidence: 99%

A robust facemask forgery detection system in video

Almukhtar¹

2022

PEN

View full text Add to dashboard Cite

show abstract

“…Presentation attack: In presentation attack, FR models are fooled by modifying the human features with the help of artifacts, such as make-up, printed face masks, 3D masks, and accessories. [23][24][25][26] As shown in Fig. 7, the presentation attack is applied using the 3D mask image and eyeglasses to fool the FR models.…”

Section: Introductionmentioning

confidence: 99%

M-SAN: a patch-based transferable adversarial attack using the multi-stack adversarial network

Agrawal

Bhatnagar

2023

J. Electron. Imag.

View full text Add to dashboard Cite

In recent times, deep neural networks (DNN) have been extensively used in several areas and have achieved great success. State-of-the-art face recognition (FR) systems have gained high accuracy using DNN. However, researchers have found that DNN-based systems fail when facing an adversarial attack on images. In adversarial attacks, the adversary modifies the face images in a manner that the human does not perceive the changes in the generated image but FR systems are unable to recognize the faces correctly. We proposed a method to generate an adversarial attack. A multistack adversarial network (M-SAN) patch-based attack is generated using the generative adversarial network under the black-box settings. The M-SAN attack targets the features of the face images using a patch to fool the FR model as targeted and untargeted attacks. In the past, several attack generation methods have been presented under white-box settings. However, white-box settings need a model architecture and information about its parameters. Due to this, a single white-box attack is not able to fool different FR models. We propose an attack generation approach that is based on black-box settings in which an attacker does not have access to the target model parameters. The attack is generated with the help of the surrogate model and then transferred to the various target models. The proposed M-SAN attack is applied to FR models including FaceNet, ArcFace, and CosFace on the labeled face in the wild dataset.

show abstract

“…Now, face recognition (FR) systems have become an exceptionally accurate technology in identifying people from images (Schroff, Kalenichenko, and Philbin 2015;He et al 2016). While being useful, face recognition may invade the privacy of individuals when used to exploit and process illicitly their face images (Hadid et al 2015;Hernandez-Ortega et al 2021) and videos (Tolosana et al 2022b,a) found on the internet, particularly social media.…”

Section: Introductionmentioning

confidence: 99%

Toward Face Biometric De-identification using Adversarial Examples

Ghafourian

Fiérrez

Gomez

et al. 2023

2023 IEEE 47th Annual Computers, Software, and Applications Conference (COMPSAC)

View full text Add to dashboard Cite

The remarkable success of face recognition (FR) has endangered the privacy of internet users particularly in social media. Recently, researchers turned to use adversarial examples as a countermeasure. In this paper, we assess the effectiveness of using two widely known adversarial methods (BIM and ILLC) for de-identifying personal images. We discovered, unlike previous claims in the literature, that it is not easy to get a high protection success rate (suppressing identification rate) with imperceptible adversarial perturbation to the human visual system. Finally, we found out that the transferability of adversarial examples is highly affected by the training parameters of the network with which they are generated.

show abstract

Introduction to Presentation Attack Detection in Face Biometrics and Recent Advances

Cited by 4 publications

References 66 publications

A robust facemask forgery detection system in video

A robust facemask forgery detection system in video

M-SAN: a patch-based transferable adversarial attack using the multi-stack adversarial network

Toward Face Biometric De-identification using Adversarial Examples

Contact Info

Product

Resources

About