|
Amina Bassit I am a postdoctoral researcher advised by Vishnu Boddeti in the Human Analysis Lab at Michigan State University (US). I received my Ph.D. degree from the University of Twente (NL) in ML-based biometric recognition and applied cryptography as part of the Marie Skłodowska-Curie funded PriMa (Privacy Matters) project, under the supervision of Raymond Veldhuis, Andreas Peter, and Florian Hahn. I won the EAB Industry Award 2024 for my Ph.D. dissertation, Fast and Accurate Biometric Search under Encryption. I am a Program Chair of the International Workshop on Biometrics and Forensics (IWBF 2025). |
|
ResearchMy research mission is to ensure that AI technologies align with core human values, preserving fundamental rights in the digital world while protecting the intellectual property of businesses. I am passionate about integrating advanced ML and applied cryptography to design scalable, secure, and privacy-preserving solutions that build trust in AI technologies for both users and service providers. |
Publications |
|
Practical Biometric Search under Encryption: Meeting the NIST Runtime Requirement without Loss of Accuracy
Amina Bassit, Florian Hahn, Raymond Veldhuis, Andreas Peter IEEE T-BIOM, 2025 Encrypted biometric search at scale becomes practical with FHE supporting SIMD and vertical chunking on the database, enabling accurate one-in-a-million search in under 10 seconds. |
|
|
Template Recovery Attack on Encrypted Face Recognition Systems with Unprotected Decision using Synthetic Faces
Amina Bassit, Florian Hahn, Zohra Rezgui, Hatef Otroshi Shahreza , Raymond Veldhuis, Andreas Peter Frontiers in Imaging, 2025 Even HE-based BTPs cannot hide from synthetic spoofs—this study shows how attackers can recover and reconstruct facial templates using only synthetic faces and a handful of comparison scores. |
|
|
A Study on the Next Generation of Digital Travel Credentials
Ana-Teodora Radutoiu, Amina Bassit, Raymond Veldhuis, Christoph Busch IEEE BIOSIG, 2024 Future digital travel credentials—tomorrow's passports—can be secured using iris biometrics, enabling clone-resistant, device-portable authentication for seamless and secure travel. |
|
|
Improved Multiplication-Free Biometric Recognition under Encryption
Amina Bassit, Florian Hahn, Raymond Veldhuis, Andreas Peter IEEE TBIOM, 2024 By turning similarity computations into lookup tables, we unlock fast and secure biometric recognition under encryption without multiplications, achieving decisions in under 50ms. |
|
|
Fast and Accurate Biometric Search Under Encryption
Amina Bassit PhD Thesis - Dec 15th, 2023 - University of Twente Lookup table-based comparisons turn biometric verification and search under encryption into milliseconds and seconds instead of minutes and hours, enabling real-world deployment of privacy-preserving biometrics without accuracy loss. Supervised by Raymond Veldhuis, Andreas Peter, and Florian Hahn Funded by the PriMa (Privacy Matters) project under the Marie Skłodowska-Curie grant No. 860315 EAB Industry Award 2024 Filled Patent jointly with the UT |
|
|
Template Recovery Attack on Homomorphically Encrypted Biometric Recognition Systems with Unprotected Threshold Comparison
Amina Bassit, Florian Hahn, Zohra Rezgui, Una Kelly, Raymond Veldhuis, Andreas Peter IEEE IJCB, 2023 Revealing comparison scores compromises BTPs' privacy. Our attack shows that a handful of fake templates and cleartext scores are enough to fully recover protected biometric templates and reconstruct raw biometric samples. |
|
|
Remote Cancelable Biometric System for Verification and Identification Applications
Hatef Otroshi Shahreza, Amina Bassit, Sébastien Marcel, Raymond Veldhuis IEEE BIOSIG, 2023 Dynamic protection meets cancelable biometrics. Our protocol enables secure, user-specific template replacement in both identification and verification without sacrificing flexibility or privacy. |
|
|
Multiplication-Free Biometric Recognition for Faster Processing under Encryption
Amina Bassit, Florian Hahn, Raymond Veldhuis, Andreas Peter IEEE IJCB, 2022 Multiplications out, lookup tables in! Our approach accelerates biometric verification under encryption without compromising recognition accuracy. Filled Patent jointly with the UT |
|
|
Hybrid Biometric Template Protection: Resolving the Agony of Choice Between Bloom Filters and Homomorphic Encryption
Amina Bassit, Florian Hahn, Raymond Veldhuis, Andreas Peter IET Biometrics, 2022 |
|
|
Transferability analysis of adversarial attacks on gender classification to face recognition: Fixed and variable attack perturbation
Zohra Rezgui, Amina Bassit, Raymond Veldhuis IET Biometrics, 2022 |
|
|
Fast and Accurate Likelihood Ratio-Based Biometric Verification Secure Against Malicious Adversaries
Amina Bassit, Florian Hahn, Joep Peeters, Tom Kevenaar, Raymond Veldhuis, Andreas Peter IEEE TIFS, 2021 |
|
|
Bloom Filter vs Homomorphic Encryption: Which approach protects the biometric data and satisfies ISO/IEC 24745?
Amina Bassit, Florian Hahn, Chris Zeinstra, Raymond Veldhuis, Andreas Peter IEEE BIOSIG, 2021 Best Paper Award |
|
|
Transferability analysis of an adversarial attack on gender classification to face recognition
Zohra Rezgui, Amina Bassit IEEE BIOSIG, 2021 |
Activities |
|
Based on Jon Barron's website. |