An application framework for evaluating methods in biometrics systems

In recent years, research in biometrics is focused on the development of new methods/algorithms to improve the accuracy of existing biometrics systems. Researchers are often required to do comparisons of their new methods/algorithms with existing/similar methods/algorithms and to make evaluations to show the performance of their biometrics systems. Currently, a universal environment in which researchers can compare and evaluate openly and objectively their biometrics methods/algorithms is rare since researchers tend to focus more on the biometrics methods/algorithms instead of providing such environment for other researchers. This paper presents Biometrika Framework: an application framework that can be used to evaluate various biometrics methods/algorithms. It is found that biometrics systems share a common topology consisting of: capturing, preprocessing, feature extraction, post-processing, matching and a data storage modules. A generic architecture of the framework is described in this paper. The framework has been tested and used to implement and evaluate two biometrics systems on fingerprint and palm-print data.