Biometric recognition systems target the automatic recognition of a person’s identity based on physical, physiological or behavioural characteristics (something a person is or produces). In particular, verification systems perform one to one authentication, comparing the captured biometric data with that person’s own biometric template (a mathematical representation of the biometric features extracted during the enrolment phase) previously stored in the system.
One of the major vulnerabilities associated to biometric recognition systems is that once a biometric template becomes compromised, it cannot be reissued, updated or destroyed. An attacker could then gain access to all the accounts/services/applications using that same biometric trait.
The main goal of the project is to create a secure multimodal biometric verification system, supporting the generation of different templates from the same biometric data, as well as cancellable templates. Furthermore, the privacy of the original biometric data needs to be guaranteed, not allowing its recovery from the stored data.
State-of-the-art methods will be extended by considering a biometric verification system based on distributed source coding principles and cryptographic hash functions, to enhance security with respect to traditional biometric systems. The additional consideration of multimodal biometrics helps preventing spoof attempts by imposters while increasing recognition rates.
|Start Date: 01-03-2010|
|End Date: 01-02-2013|
|Team: Paulo Luis Serras Lobato Correia, Luís Eduardo de Pinho Ducla Soares, Maurício Abreu Ramalho, Sandro Noto|
|Groups: Multimedia Signal Processing – Lx|
|Local Coordinator: Paulo Luis Serras Lobato Correia|