Abstract
Biometric authentication is, over time, becoming an indispensable complementary component to traditional authentication methods that use passwords and tokens. As a result, the research interest in the protection techniques for the biometric template has also grown considerably. In this paper, we present a light-weight AI-based biometric authentication that operates based on the binary representation of a biometric instance. In details, a binary classifier will be trained using the binary strings that represent the intraclass and interclass biometric subjects. The Support Vector Machine and Multi-layer Perceptron Neural Network are chosen as the classifier to evaluate the fingerprint-based and iris-based authentication capability. Afterward, the authenticated biometric string is fed to a hash function to produce a hash value, which is to be used in a Zero-Knowledge-Proof Protocol for the purpose of privacy preservation. In order to improve the recognition of the classifier, we devise a simple yet efficient strategy to enhance the discriminativeness of the binary strings and name it the Composite Features Retrieval. We evaluated the proposed method with the four publicly available fingerprint datasets FVC2002-DB1, FVC2002-DB2, FVC2002-DB3, and FVC2004-DB2 and the iris dataset UBIRISv1. The promising performance shows this method's capability.
Highlights
Biometric authentication systems have blossomed over the last few years as the mobile devices are becoming more and more popular
In order to evaluate the performance of the classifiers used, we use the False Acceptance Rate (FAR), False Rejection Rate (FRR), and Accuracy
FAR is the probability that the model mistakenly accepts a sample that is not from the same subject with which the model was trained
Summary
Biometric authentication systems have blossomed over the last few years as the mobile devices are becoming more and more popular. Biometric authentication has been widely deployed in mobile devices such as mobile phones and laptops which are lost or get stolen Once such device is in the wrong hands, there exist many adversary attacks that can retrieve sensitive information or data stored in the platform via side channel attacks [1]. There exist two issues with this framework: (1) Biometrics query samples fed into the SVM classifier need to go through the error correction code (ECC). The proposed scheme does not distort the original biometrics feature distribution that is fed into the SVM classifier and does not store any secret in the device either. The proposed scheme does not distort the original biometric feature distribution that is fed into the classifier and does not store any secret in the device either. Reviews the related work in the field; the details of our method are presented in section III while experimental results are shown in section IV; section V concludes the paper
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 callDisclaimer: 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.
