Something about Biometrics
PRA Lab has been working on biometric recognition systems since 2001. The so-called biometrics are physiological (fingerprint, face, iris, retina, palmprint) or behavioural (whilst voice, gait) human characteristics.
Fingerprint classification consists in classifying the shape given by the flow of ridges and valleys on a fingerprint among the five fundamental classes defined by Sir Edward Henry in 1890.
Multiple biometric systems for personal recognition
PRA Lab concurred to the state-of-the-art advancement on multi-biometric systems based on fingerprint and face biometric traits. Our main contributions consist of developing novel algorithms and frameworks for combining information coming from different biometric traits, and soft-biometrics.
Adaptive biometric systems based on template update
Biometrics are subject to temporal variations so, a "template" acquired at a certain time could be no more representative of the subject’s identity. PRA Lab developed several algorithms aimed to automatic template update. Such algorithms have been developed for both mono- and multimodal systems.
Fingerprint and face liveness detection
Fingerprints can be replicated using artificial materials. PRA Lab has developed a high-level skill on replicating fingerprints by very common materials and has developed several algorithms for detecting the fingerprint liveness which are competitive with those in the literature.