Multimodal Biometrics Recognition by Dimensionality Reduction Method

Multimodal biometric system utilizes two or moreindividual modalities, e.g., face, ear, and fingerprint, toimprove the recognition accuracy of conventional unimodalmethods. We propose a new dimensionality reduction methodcalled Dimension Reduce Projection (DRP) in this paper. DRPcan not only preserve local information by capturing the intramodal geometry, but also extract between-class relevant structures for classification effectively. Experimental results show that our proposed method performs better than other algorithms including PCA, LDA and MFA.