Dynamic signature verification and the human biometric sensor interaction model

Abstract

This paper examines the application of the HBSI model to a range of dynamic signature verification capture digitizers. Behavioral biometrics (which considers the process by which a human performs a function) are inherently more complicated to analyze as they may contain a temporal modification as part of a valid capture interaction. In this study, a framework for the development of an HBSI model is outlined, based on two different signature digitizers. Both devices enabled the capture of both temporal data (dynamic) as well as an image of the completed (static) signature. Using this data, we propose a signature modality revision to the HBSI model with interaction errors classified into five categories. This study deals solely with genuine users, examining both correct and incorrect signature presentations.

Biometric Systems Interaction Assessment: The State of the Art

“…Works in this area aim to improve ergonomics and produce the most appropriate feedback for users, therefore reducing FTA. One of the first specific works in this area is the integration of dynamic signature verification with the HBSI model [26] (see Section II-D). The authors revealed the complexity of the potential interactions according to the scenario (signing an important contract, in the supermarket, etc.).…”

Exploring Mobile Biometric Performance Through Identification of Core Factors and Relationships

“…The development of the Human-Biometric Sensor Interaction (HBSI) model [26] investigated how scenarios can modify performance. Brockly et al [27] concluded "the development [of HBSI] reveals the complexity of the potential interactions and the changes of those interactions when digitisers change, as well as when the ceremony changes".…”

Automatic usability and stress analysis in mobile biometrics

“…One of the most complete models published up to now is the Human Biometric System Interaction (HBSI) [4], which proposes methods and measures (including the ones recommended by ISO 13407:1999) to analyse the user-sensor interaction deeply. As this model has not yet been tested empirically in dynamic modalities [5], our work goes a step further, proposing some modifications to it, and therefore, the results obtained can be considered a novelty. Furthermore, this study includes stress tests where users sign under pressure conditions.…”