Decision Level Fusion in Standardized Fingerprint Match-on-Card

Abstract

Biometric user authentication is becoming popular for a variety of applications. Fingerprint verification is one of the most reliable and well-understood biometric measures. The match-on-card technology allows a high level of accuracy and privacy for biometric systems. Open systems adhere to interindustry standards while proprietary implementations can take advantage of rich information contained in biometric traits such as fingerprints. The trade-off between accuracy and interoperability needs to be resolved. We designed a hybrid match-on-card algorithm that combines standardized templates from two different standards to make up a system with enhanced matching accuracy.

An Approach to Biometric Identity Management Using Low Cost Equipment

“…The algorithmic approach shown in Fig. 4 is similar to the detailed description in [3] except that the additional preprocessing is required to cope with the webcam generated fingerprint images. StarSign Mobility Token MicroSD hardware, picture taken from [16] VI.…”

“…For details on how to generate this data, [3] is suggested reading. Raw webcam image, processed and binarized image, extracted minutiae and skeleton Conversely to most common optical and silicon fingerprint sensors, light pixels of the fingertip represent ridges while darker pixels indicate valleys.…”

“…This product includes a smart card chip with an ISOcompliant On-card-biometric-comparison implementation described in [3]. Fig.…”

“…Normally, these features are fingerprint minutiae. We decided to use the hybrid algorithm introduced in [3], which relies on minutiae and skeleton data. Fig.…”

An evaluation of indirect attacks and countermeasures in fingerprint verification systems

“…This is a good method to protect the privacy of users (their templates do not leave the card), while providing reasonable performance with current technology. Some related works on fingerprint MoC have been reported by Bistarelli et al (2006) and Mueller and Martini (2006). In the experiments the NIST software is used in the feature extraction process and the resulting templates are transformed to the MoC system format and sent to the smart card.…”