An Effective Multimodal Biometrics Recognition System using Hybrid Fusion Technique for Efficient Authentication
Neeru Bala, Anil Kumar, Rashmi Gupta · 2022 8th International Conference on Signal Processing and Communication (ICSC) · 2022
Biometric system has been the topic of research form several decades in order to verify and identify an individual. To alleviate downsides of unimodal biometric system which relies on one trait, integration of several cues is a logical strategy. A multimodal biometric system employs multiple traits for authentication of a person. In this work hybrid level of fusion of iris, palmprint and finger-knuckle-print has been proposed. In this work to achieve hybrid level of fusion of the traits firstly same feature descriptor is employed to extract the features of all the three and then feature vectors are concatenated and match scores are obtained. The matching scores of feature level fusion techniques are fused with the matching scores of finest unimodal biometric system. The relative evaluation of anticipated approach with single level fusion multimodal system reveals that proposed approach outshines in terms of all the performance metrics. State-of-the-art comparisons further support the suggested approach’s superior performance.