Biometric Menagerie in Time-Span Separated Fingerprint Data
Simon Kirchgasser, Andreas Uhl · 2016
Multiple factors are influencing the performance of fingerprint recognition systems. Some of those depend on the used recognition implementations or datasets, while other factors like fingerprint ageing can be difficult to isolate. The aim of this research is the consideration of user related characteristics which have been introduced as the so called "Doddington's Zoo" to describe possible present template ageing influence. Certain user dependent weaknesses could be influenced by fingerprint ageing such that those system vulnerabilities are amplified or even attenuated. To investigate this aspect, the users in the databases (including a time separation of 4 years) are labelled according to the main model provided by the menagerie concept. The analysis of the single categories revealed that the animal groups are not extended (labelled users in the older datasets are not the same as in the newer ones) regardless which dataset and recognition system is considered.