Multimodal biometrics by fusion for security using genetic algorithm

The physiological biometrics like face is combined with behavioral biometrics like speech to achieve the robustness of fusion process of a multimodal system. The selection of the biometrics is dependent on the robustness and uniqueness of the biometric. That is why, the selection of these two biometrics is done in this work. Mel Frequency Cepstral Coefficients has been utilized for speech feature extraction and in addition to this fuzzy logic is also utilized for training purpose. Then, the optimized features values are reduced using genetic algorithm. In the end, fusion is achieved by combination of fuse values obtained from both 2 biometrics. The whole simulation is tested in MATLAB environment.