Publication

Utilization of optical image processing techniques on Biometrics for privacy preservation

Nassar, R.M
Khalaf, A.A.
El-Banby, G.M
El-Samie, Fathi E. Abd
El-Shafai, Walid
El-Gazar, Safaa
Research Projects
Organizational Units
Journal Issue
Abstract
Cancelable biometric systems represent a new trend to solve traditional access problems. Generally, traditional access using passwords and token cards has many problems. The most important problem is the ease of loss and counterfeiting. The main advantage of biometrics in the authentication process is the ease of implementation of biometric systems, as biometric traits always accompany persons without the need for either mental or manual intervention. A single biometric template can be used for identification, but the system in this case is still subject to different types of attacks. In addition, the biometric traits may be lost in hacking scenarios. To reduce the risk of losing biometric traits, the trend of cancelable biometrics has been developed based on distorted or encrypted versions of the biometric templates for authentication. In addition, merging more than one biometric template is another solution for a reliable authentication system. A survey of unimodal and multimodal biometric systems is presented in this paper. Furthermore, a comparison study of previous approaches is presented. Several algorithms are used to merge biometrics in multimodal cancelable biometric systems, such as Discrete Cosine Transform (DCT) and Singular Value Decomposition (SVD). Optical encryption and lossy compression are also used as tools to generate the cancelable templates from the merged biometrics. The Equal Error Rate (EER) and the Receiver Operating Characteristic (ROC) curve are the most popular metrics used to evaluate cancelable biometric systems. The comparison study indicates that the DCT-based systems can achieve an EER close to 0 and an Area under the Receiver Operating Characteristic curve (AROC) close to 1.
Publisher
Sponsor
Copyright
Book title
Journal title
Embedded videos