ATHENE's diverse contributions showcase the company's proficiency in critical areas of biometric security.

The accepted papers are:
StableMorph: High-Quality Face Morph Generation with Stable Diffusion
Authors: Wassim Kabbani, Kiran Raja, Raghavendra Ramachandra, Christoph Busch
The researchers are developing a system that uses stable diffusion technology to generate high-quality face morphs. This method enables the creation of realistic morphed faces for the security testing of biometric systems.

Can foundation models predict fitness for duty?
Authors: Juan Tapia, Christoph Busch
This study investigates whether foundation models can predict a person's fitness for duty based on biometric data. The paper analyses the limitations and ethical implications of such AI-based assessment systems.

Closing the Performance Gap in Biometric Cryptosystems: A Deeper Analysis of Unlinkable Fuzzy Vaults
Authors: Hans Geißner, Christian Rathgeb
The authors analyse performance deficits in biometric cryptosystems, with a focus on unlinkable fuzzy vaults. Their research aims to bridge the security gap between theoretical models and practical implementations.

Training-free dimensionality reduction via feature truncation
Authors: Florian Bayer, Maximilian Russo, Christian Rathgeb
This paper presents a training-free method for dimensionality reduction via feature truncation. This method improves the efficiency of privacy-preserving multi-biometric systems by eliminating the need for complex learning processes.

Trade-offs in cross-domain generalization of foundation models fine-tuned for biometric applications
Authors: Tahar Chettaoui, Naser Damer, Fadi Boutros
The researchers investigate the trade-offs involved in the cross-domain generalisation of foundation models for use in biometric applications. The focus is on optimising fine-tuning strategies for different biometric modalities.

LivDet2025: Towards Robust and Generalisable Fingerprint Presentation Attack Detection
Author: Giulia Orru et al.
This paper outlines a competition to detect fingerprint presentation attacks robustly and generally. The aim is to develop reliable algorithms that can cope with unknown attack types.

SynID: Passport Synthetic Dataset for Presentation Attack Detection
Authors: Juan Tapia, Fabian Stockhardt, Lázaro Janier González-Soler, Christoph Busch
The authors are creating a synthetic dataset of passport documents for research on presentation attacks. This dataset enables the development and evaluation of detection systems without the use of real ID documents.

Iris Liveness Detection Competition (LivDet-Iris) 2025
Author: Adam Czajka et al.
This competition focuses on liveness detection in iris-based biometric systems. Participants develop algorithms to distinguish between living eyes and various types of spoofing attack.

Second Competition on Presentation Attack Detection on ID Card
Juan Tapia et al.
The researchers tested algorithms to detect fake ID cards.

ATHENE researchers are also organising a special session entitled 'Recent Advances in Detecting Manipulation Attacks on Biometric Systems (ADMA-2025)', as well as two major competitions: the 'Second Competition on Presentation Attack Detection on ID Cards' and 'LivDet2025: Towards Robust and Generalisable Fingerprint Presentation Attack Detection'.

More about the conference

show all news