Multimodal Biometric System based Face-Iris Feature Level Fusion

Abstract

This paper proposed feature level fusion technique to develop a robust multimodal human identification system. The humane face-iris traits are fused together to improve system accuracy in recognizing 40 persons taken from ORL and CASIA-V1 database. Also, low quality iris images of MMU-1 database are considered in this proposal for further test of recognition accuracy. The face-iris features are extracted using four comparative methods. The texture analysis methods like Gray Level Co-occurrence Matrix (GLCM) and Local Binary Pattern (LBP) are both gained 100% accuracy rate, while the Principle Component Analysis (PCA) and Fourier Descriptors (FDs) methods achieved 97.5% accuracy rate only.