查看文章

Multibiometric systems are a promising area that addresses a number of unimodal biometric systems drawbacks. The main limit of these systems is the lack of information in terms of quantity (number of discriminant features) and quality (diversity of information, correlation…). Using multiple sources of information and/or treatment is a solution to overcome these problems and enhance system performances. Performance requirements of current systems related to context use involve designed solutions that optimally satisfy security requirements. This can represent an optimization problem that aims at searching the optimal solution matching security needs. In our study, we are interested in combining different score level rules using an evolutionary algorithm. We use Genetic Algorithm to derive a score fusion function based on primitive operations. The process uses an optimized tree to determine function …