Abstract:

Abstract Multi-modal biometric system combines feature knowledge indifferent traits to conquer the deficienciesof unimodal systems. But, over the last few years, the conventional biometric system focuses on the usage of handcrafted feature for human recognition. Feature extraction is certainly a crucial phase for the performance of this approach, because of the complexityofmounting consistent features to handle variations in the provided image. This article introduces an Optimal Machine Learning-based Automated Multimodal Biometrics Recognition for Person Identification (OML-AMBRPI) technique. The presented OML-AMBRPI technique makes use of three different biometrics namely iris, fingerprint, and microarray images for person recognition. For feature extraction, the OML-AMBRPI technique uses eXtendedCenter-Symmetric Local Binary Pattern (XCS-LBP) and Gabor feature extraction. Next, feature fusion process is carried out by the decision level fusion technique. At last, the multi-objective grey wolf optimizer (MOGWO) with feed forward neural network (FFNN) for the recognition process. The experimental outcome analysis of the OML-AMBRPI system demonstrates the significant performance of the OML-AMBRPI technique over other models.

Keywords: