Face and Thumb Based Multimodal Bio-Metric Authentication using Harris Feature Extraction and Stenography
Pallavi S Biradar1, Anand Jatti2

1Pallavi S Biradar, M.Tech Scholar, Dept. Of Electronics and Instrumentation Engineering, RV College of Engineering, Bangalore India.
2Dr. Anand Jatti, Associate Professor, Dept. Of Electronics and Instrumentation Engineering, RV College of Engineering, Bangalore India.

Manuscript received on August 01, 2020. | Revised Manuscript received on August 05, 2020. | Manuscript published on September 30, 2020. | PP: 550-555 | Volume-9 Issue-3, September 2020. | Retrieval Number: 100.1/ijrte.C4629099320 | DOI: 10.35940/ijrte.C4629.099320
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: In the past recent, identification of a person in an effective manner is a foremost concern for any security authentication in numerous applications such as, banking, e-commerce, communications etc. One of the best identification technology for person identification and authentication compared with the existing password based authentication is the multimodal biometric technology. Multimodal can be defined as, a system which uses two or more biometrics for identification of person. In the paper we propose a multimodal bio-metric system with a unique methodology and features extraction method incorporated in system for a secure authentication. The two modalities used in the system are face and thumb. We use Harris based image feature extraction for both and choose the best unique features from both and fused using concatenation. The extracted unique features are embedded in a cover image using modulo operator based steganography technique. This encrypted data is shared as an image file to the receiver for authentication. At the receiver end the hidden features are decrypted and separated into face and thumb features. These decrypted features are compared with the pre-trained authorized person feature, based on the multi-svm classifier result the person is decided as authorized or unauthorized. The accuracy of the system is been calculated and was resulted in a good accuracy. The system can be made much more secure by adding an additional secret key for encryption and decryption. 
Keywords: Bio-metric, Harris, modulo operator, multi-modal, multi-svm. Stenography.