Ahmed Mahfouz

Dr. Ahmed Mahfouz

orcid

Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
14 September 2024 in Neural Computing and Applications
Verified
Journal Article
Closed access
PDF file
AMAhmed Mahfouz
Ahmed Mahfouz
HMHebatollah Mostafa
Hebatollah Mostafa
TMTarek M Mahmoud
Tarek M Mahmoud
AEAhmed Sharaf Eldin
Ahmed Sharaf Eldin

22 August 2025 in Scientific Reports
Verified
Journal Article
Open access
PDF file
SKSohair Kilany
Sohair Kilany
AMAhmed Mahfouz
Ahmed Mahfouz
Face Verification (FV) systems have exhibited remarkable performance in verification tasks and have consequently garnered extensive adoption across various applications, from identity duplication to authentication in mobile payments. However, the surge in popularity of face verification has raised concerns about potential vulnerabilities in the face of adversarial attacks. These concerns originate from the fact that advanced FV systems, which rely on deep neural networks, have recently demonstrated susceptibility to crafted input samples known as adversarial examples. Although imperceptible to human observers, adversarial examples can deceive deep neural networks during the testing and deployment phases. These vulnerabilities raised significant concerns about the deployment of deep neural networks in safety-critical contexts, prompting extensive investigations into adversarial attacks and corresponding defense strategies. This comprehensive survey provides a comprehensive overview of recent advances in deep face verification, encompassing a broad spectrum of topics such as algorithmic designs, database utilization, protocols, and application scenarios. Furthermore, we conduct an in-depth examination of state-of-the-art algorithms to generate adversarial examples and the defense mechanisms devised to mitigate such adversarial threats.

1 January 2023 in Computers, Materials & Continua
Verified
Journal Article
Open access
PDF file
ASAwny Sayed
Awny Sayed
SKSohair Kinlany
Sohair Kinlany
AZAlaa Zaki
Alaa Zaki
AMAhmed Mahfouz
Ahmed Mahfouz
Face verification systems are critical in a wide range of applications, such as security systems and biometric authentication. However, these systems are vulnerable to adversarial attacks, which can significantly compromise their accuracy and reliability. Adversarial attacks are designed to deceive the face verification system by adding subtle perturbations to the input images. These perturbations can be imperceptible to the human eye but can cause the system to misclassify or fail to recognize the person in the image. To address this issue, we propose a novel system called VeriFace that comprises two defense mechanisms, adversarial detection, and adversarial removal. The first mechanism, adversarial detection, is designed to identify whether an input image has been subjected to adversarial perturbations. The second mechanism, adversarial removal, is designed to remove these perturbations from the input image to ensure the face verification system can accurately recognize the person in the image. To evaluate the effectiveness of the VeriFace system, we conducted experiments on different types of adversarial attacks using the Labelled Faces in the Wild (LFW) dataset. Our results show that the VeriFace adversarial detector can accurately identify adversarial images with a high detection accuracy of 100%. Additionally, our proposed VeriFace adversarial removal method has a significantly lower attack success rate of 6.5% compared to state-of-the-art removal methods.

28 January 2018 in arXiv
Verified
Preprint
Open access
PDF file
AMAhmed Mahfouz
Ahmed Mahfouz
TMTarek M Mahmoud
Tarek M Mahmoud
AEAhmed Sharaf Eldin
Ahmed Sharaf Eldin
Recent research has shown the possibility of using smartphones' sensors and accessories to extract some behavioral attributes such as touch dynamics, keystroke dynamics and gait recognition. These attributes are known as behavioral biometrics and could be used to verify or identify users implicitly and continuously on smartphones. The authentication systems that have been built based on these behavioral biometric traits are known as active or continuous authentication systems. This paper provides a review of the active authentication systems. We present the components and the operating process of the active authentication systems in general, followed by an overview of the state-of-the-art behavioral biometric traits that used to develop an active authentication systems and their evaluation on smartphones. We discuss the issues, strengths and limitations that associated with each behavioral biometric trait. Also, we introduce a comparative summary between them. Finally, challenges and open research problems are presented in this research field.

10 September 2026 in Mobile Networks and Applications
Verified
Journal Article
Closed access
PDF file
AMAhmed Mahfouz
Ahmed Mahfouz
KAKamal Almouh
Kamal Almouh
KSKhouloud Samrouth
Khouloud Samrouth
NBNader Yasser Bakir
Nader Yasser Bakir
GCGerard Chalhoub
Gerard Chalhoub

22 June 2026 in Future Internet
Verified
Journal Article
Open access
PDF file
AMAhmed Mahfouz
Ahmed Mahfouz
MHMohammed Abdulla Salim Al Husaini
Mohammed Abdulla Salim Al Husaini
Alaa A. K. IsmaeelAlaa A. K. Ismaeel
Alaa A. K. Ismaeel
YHYousuf Nasser Said Al Husaini
Yousuf Nasser Said Al Husaini
A single user-invariant tensor, kinematically impossible for any human finger to produce, bypasses bidirectional long short-term memory (BiLSTM) continuous-authentication defenders with numerically identical structure across four independently trained generators. We arrive at this finding by training generative adversarial networks against BiLSTM defenders on 51 users across three independent random seeds, with the data partition held fixed, to test the prevailing assumption that successful generative attacks must reproduce the victim’s kinematic behavior. Aggregate attack success rate varies from 31.4% to 45.1% across seeds, a 13.7 percentage-point spread arising purely from optimization stochasticity, demonstrating how unreliable single-seed reporting is as an estimator of the true attack surface. A four-group descriptive stratification shows that 8% of users are attacked across all three seeds, 31% are consistently safe, and 61% exhibit seed-dependent outcomes. Classifier accuracy on zero-effort impostors does not predict adversarial vulnerability (Spearman ρ=−0.058, permutation p=0.688), whereas intra-user behavioral variance does (ρ=+0.351, permutation p=0.012, Bonferroni-corrected). The mechanism is not behavioral emulation but convergence to an Adversarial Skeleton Key, a tensor located in an unregularized region of the BiLSTM’s decision surface that the network reliably maps to acceptance, despite lying many standard deviations outside any genuine human distribution. The mimicry-centric evaluation paradigm underestimates the real threat surface. Input-space plausibility must be treated as a defensive layer rather than a preprocessing concern.