A New Multimodal Recognition Technique Without Subject's Cooperation Using Neural Network Based Self Organizing Maps

摘要

In this manuscript we present a new multimodal biometric system based on neural networks self organizing maps (SOM) for the detection and recognition of face, ear and hand geometry. We use combined principal component analysis (PCA) and SOMs for the dimensionality reduction and then use it for the combined search space optimization of ear, face and hand geometry. We name our method named RJSOM. We show that the proposed RJSOM method improves the performance and robustness of recognition when compared to methods proposed in literature. We apply the proposed method to a variety of datasets and show the results.