Abstract
In this paper, a novel fusion method based on Total Error Rate (TER) and multiple hidden layer probabilistic extreme learning machine is proposed. At first, the study transfers the matching scores into TER based on corresponding False Reject Rates (FRR) and False Accept Rates (FAR) aims at avoiding to calculating the posterior probability. At the second, a new fusion strategy based on multiple hidden layer probabilistic extreme learning machine is introduced, which optimizes the architecture of hidden nodes by weighted calculation of different output matrices and then transforms the numeric output of ELM to the probabilistic outputs and unifies the outputs in a fixed range, the matrices weights and the output weights are optimized using a hybrid intelligent algorithm based on differential evolution and particle swarm optimization. Experiment result shown that the proposed method renders very good performance as it is quite computationally and outperforms the traditional score level fusion schemes, the experimental result also confirms the effectiveness of the proposed method to improve the performance of multibiometric system.
Highlights
Face, fingerprint and iris have been explored to recognize humans in some extent[1]
Compared to the unimodal recognition system, the multimodal biometric system have several merits:1) it is address the problem of nonuniversality which encountered by unimodal system; 2)they limit the ability of an impostor to spoof multiple biometric features of a legitimately enrolled sample;3) it is can solve the problem of noisy input data effectively;4) it can viewed as fault tolerant recognition systems due to the reason of they can operate continuously even when some biometric resources become unbelievable and unreliable
A new fusion strategy based on hybrid intelligent multiple hidden layer probabilistic extreme learning machine(HMP-ELM) is introduced, which optimizes the architecture of hidden nodes by weighted calculation of different output matrices and transforms the numeric output of ELM to the probabilistic outputs and unifies the outputs in a fixed range, in this strategy, the matrices weights μ of the Multiple Hidden Layer Extreme Learning Machine(M-ELM) and the output weights β belongs to the ELM are optimized using a hybrid intelligent algorithm based on Differential Evolution and Particle Swarm Optimization(DEPSO)
Summary
Fingerprint and iris have been explored to recognize humans in some extent[1]. the recognition performance of these unimodal biometric recognition system are not ideal in some case because they are plagued by some drawbacks. Confronted by the limitation of the unimodal recognition systems, multimodal biometric recognition systems try to alleviate these disadvantages by consolidating the evidence presented by multiple biometric sources in order to in order to improve the recognition performance. Compared to the unimodal recognition system, the multimodal biometric system have several merits:1) it is address the problem of nonuniversality which encountered by unimodal system; 2)they limit the ability of an impostor to spoof multiple biometric features of a legitimately enrolled sample;3) it is can solve the problem of noisy input data effectively;4) it can viewed as fault tolerant recognition systems due to the reason of they can operate continuously even when some biometric resources become unbelievable and unreliable. Score level fusion is the most frequently utilized because of easy availability of the scores and contains ample information to discriminate between genuine and impostor scores. Already existing score level fusion approaches can be categorized into three classes: transformation-based, density-based and classifier-based
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: International Journal of Computational Intelligence Systems
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
