Steinbeis Transfer Center BISE - Biometrics

Consulting

• Biometrics
• Image processing
• Machine learning
• IT security
• Data protection
• Standardisation

• BTAP: The Biometric Transaction Authentication Protocol (BTAP)
  The Biometric Transaction Authentication Protocol (BTAP) addresses a fundamental problem in today’s online banking: verifying the authenticity and integrity of personal and transaction data during financial transactions in a potentially insecure environment. Furthermore, BTAP fulfils the requirement to use biometric data in online banking scenarios to establish a strong link between biometric characteristics and transaction data. At the same time, sensitive biometric information is sealed using a privacy-friendly mechanism and cannot be extracted from the protocol flow by any other party. BTAP offers two key elements of a security protocol: low complexity and strong modularisation capabilities.  
• MAD: Morphing Attack Detection
  Recent research has shown that attacks targeting the merging of photographs (so-called morphing attacks) pose a serious security risk to facial recognition systems. Reliable detection of a morphing attack remains a major challenge for research, as the cameras and morphing techniques used by an attacker are unknown. This paper presents a robust Morphing Attack Detection (MAD) system that can also generalise across database boundaries.
• SOTAMD: State of the Art in Morphing Detection
  The aim of the SOTAMD project is to determine the state of the art in mechanisms for detecting morphed facial images by collecting a dataset of morphed facial images from various locations and capture systems, for which image quality is guaranteed in accordance with ICAO and EU Regulation 2252/2004. Furthermore, the project will evaluate existing mechanisms for detecting morphed facial images based on a common, established benchmark protocol and metrics for morphed facial image detection. The project focuses on the difference testing scenario for morphing detection, in which a bona fide facial image is checked against the target image (e.g. the printed image on a passport application) in a signal comparison. The project addresses the scenario of automated border control, in which the live facial image (captured at the eGate) is compared with the image stored on the passport chip.
• BioBida: Biometrics and Big Data
  In view of the current growth and benefits of biometric systems, this project investigates methods that support the effective and accurate processing of large volumes of biometric data. The aim of this project is to develop efficient algorithms and data structures for biometric identification that can deliver search results on very large biometric datasets in real time whilst ensuring the protection of sensitive biometric data.
• RESPECT: Reliable, secure and privacy-preserving multi-biometric personal authentication
  This Franco-German collaborative project investigates the potential of multi-biometric approaches as a means of defending against various presentation-of-self attacks and improving generalisation capabilities, whilst ensuring the protection of sensitive data. The research focuses on the use of biometric features that can be easily and reliably captured using ubiquitous mobile devices, and biometric features that enable computationally manageable homomorphic encryption. The project’s objectives encompass both the robustness and security of biometric capture and data protection mechanisms. At the same time, the aim is to improve recognition accuracy and reliability, thereby strengthening confidence in future multi-biometric audiovisual recognition technologies and realising their societal and economic benefits.