Feature-Level vs. Score-Level Fusion in the Human Identification System

The design of a robust human identification system is in high demand most modern applications such as internet banking and security, where the multifeature biometric system, also called feature fusion one common solutions that increases reliability improves recognition accuracy. This paper implements comprehensive comparison between two methods, named feature-level score-level fusion, to determine which method highly overall performance. takes into consideration image quality for six combination datasets well type applied extraction method. four local binary pattern (LBP), gray-level co-occurrence matrix (GLCM), principle component analysis (PCA), Fourier descriptors (FDs), are separately generate face-iris machine vector dataset. experimental results highlighted accuracy has been significantly improved when texture descriptor method, LBP, or statistical PCA, utilized with rather than all datasets. maximum obtained at 97.53% LBP Euclidean distance (ED) considered measure rate minimum equal error (EER) value.