Multimodal Biometric Recognition: Fusion of Modified Adaptive Bilinear Interpolation Data Samples of Face and Signature using Local Binary Pattern Features
Arjun B C1, H N Prakash2
1Arjun B C, Information Science & Engineering Department, Rajeev Institute of Technology affiliated to Visvesvaraya Technological University, Belgaum, Karnataka,India.
2H N Prakash, Computer Science & Engineering Department, Rajeev Institute of Technology affiliated to Visvesvaraya Technological University, Belgaum, Karnataka,India.
Manuscript received on January 26, 2020. | Revised Manuscript received on February 05, 2020. | Manuscript published on February 30, 2020. | PP: 3111-3120 | Volume-9 Issue-3, February 2020. | Retrieval Number: C6117029320/2020©BEIESP | DOI: 10.35940/ijeat.C6117.029320
Open Access | Ethics and Policies | Cite | Mendeley
© 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: Biometric based authentication systems use particular person characteristics which might be based on either behavior like voice, signature etc. or body structure like face, iris, palm print, fingerprint, etc. The performance of any unimodal biometric arrangement is depending on elements like surroundings, atmosphere, and sensor precision. Also, there are numerous trait unique demanding situations which include pose, expression, growing old and so forth for face reputation, occlusion and acquisition related problems for iris and terrible high-quality and social popularity related troubles for fingerprint. Hence, fusion of more than one biometric samples, traits or algorithms to achieve quality performance is another way to reap the better overall performance. In current scenario many researchers concentrating on Multimodal Biometrics with new fusion techniques ideas. We propose a new method of feature level fusion which uses Modified Adaptive Bilinear Interpolation (MABI) method to increase the resolution of data sample, which gives better features for fusion which gives more accurate results. In this work, experiment is done on AT&T face Cambridge University Computer Laboratory and MCYT signature Biometric Recognition Group datasets with combination of both unimodal and multimodal traits. K Nearest Neighbor (KNN) and Ensemble methods are used for classification. The proposed biometric system can be used in biometric surveillance, biometric screening for secured places, forensic applications etc.
Keywords: Biometrics; Unimodal biometric; Multimodal biometric; Local Binary Pattern (LBP), Feature Level Fusion; Modified Adaptive Bilinear Interpolation (MABI).