ON THE EFFECT OF SCORE EQUALIZATION IN SVM MULTIMODAL BIOMETRIC SYSTEMS

摘要: Most Support Vector Machine (SVM) based systems make use of conventional methods for the normalization features or scores previously to fusion stage. In this work, in addition methods, two equalization histogram equalization, which was recently introduced multimodal systems, and Bi-Gaussian is presented paper, are applied upon a person verification system composed by prosodic, speech spectrum, face information. The techniques have obtained best results; concretely, outperforms more than 22.19 % results Min-Max normalization, most used technique SVM systems. prosodic spectrum been provided experts using records Switchboard I database recognition XM2VTS database.