A Robust Multimodal Biometric System VIA Multiple Svms
Komal1, Chander Kant2
1Ms. Komal, Research Scholar, Department of Computer Science & Applications, K.U., Kurukshetra (Haryana), India.
2Dr. Chander Kant, Assistant Professor, Department of Computer Science & Applications, K.U., Kurukshetra (Haryana), India.
Manuscript received on 09 September 2019 | Revised Manuscript received on 18 September 2019 | Manuscript Published on 11 October 2019 | PP: 203-209 | Volume-8 Issue-11S September 2019 | Retrieval Number: K104209811S19/2019©BEIESP | DOI: 10.35940/ijitee.K1042.09811S19
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open-access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Generally single Support Vector Machine (SVM) is employed in existing multimodal biometric authentication techniques, and it assumes that whole set of the classifiers is available. But sometimes it is not possible due to some circumstances e.g. injury, some medical treatment etc. This paper includes a robust multimodal biometric authentication system that integrates FKP (Finger-Knuckle Print), face and finger-print at matching score level fusion using multiple parallel Support Vector Machines (SVMs). Multiple SVMs are applied to overcome the problem of missing biometric modality. Every possible combination of three modalities (FKP, face and finger-print) are taken into consideration and all combinations have a corresponding SVM to fuse the matching scores and produce the final score set for decision making. Proposed system is more flexible and robust as compared to existing multimodal biometric system with single SVM. The average accuracy of proposed system is estimated on a publicly available dataset with the use of MUBI tool(Multimodal Biometrics Integration tool) and MATLAB 2017b.
Keywords: Face Recognition, Finger Knuckle Print Recognition, Finger Print Recognition, Support Vector Machines (SVMs).
Scope of the Article: Pattern Recognition