A feature fusion scheme for multi-biometric template protec

In multi-biometric systems there are multiple templates generated for each subject. Every template can use different schemes according to the biometric characteristic types it represents and also according to the requirements of the system. These templates can be fused and processed further. Fusion can happen at: feature level; score level; or decision level. Fusion at feature level means that features obtained from different feature extractors are used to generate a vector and then all the vectors can be concatenated into a single one.In general fusion techniques treat the different biometric characteristic types as different statistical tests and then combine the results. In our paper we present a fusion scheme which considers different biometric data and stores them in a matrix which is then converted to an image. These data will be normalized and will be weighted to generate a number which will serve as one of the dimensions of the matrix.