Biometrics & Biostatistics

Biography

Biography: Ramón Santana Fernández

Abstract

The implementation of biometrics solutions for users’ authentication in daily tasks has caused great concern about the safety and privacy of the biometric data. Different vulnerabilities detected on automated fingerprint identification systems could expose the minutiae if the templates are in plain text. To solve this security issue several minutiae template protection models have been proposed like fuzzy vault, biohashing and cancelable templates; however the minutiae alignment process is required before the template matching is executed to increase the probabilities to find a positive match. In order to protect efficiently the biometric data it is necessary to meet three basic requirements: Cryptographic security, revocability and performance; however, most of the models described to date fail in this task. The fingerprint minutiae template protection scheme must capture as much identifying information of the fingerprint as possible and solve the problem of the template alignment before the matching process is executed in the protected domain. A study on the fingerprint minutiae template protection models, specifically those that start the process from features derived of the minutiae using minutiae structures, their main strengths, weaknesses and vulnerabilities was conducted in this work. Analyzing the types of attacks described in the bibliography to obtain the original biometric data from protected templates and the attacks performed to the minutiae triplets like minutiae vicinity decomposition is the main objective of this research. As a result, the vulnerabilities of each minutiae structure are identified, the elements to propose a new minutiae structure are analyzed and initial results are discussed.