BMT 2015 : Special Issue: Biometric Performance and Statistics

Special Issue: Biometric Performance and Statistics <br> <br>Biometrics technology increasingly plays a central role in personal, national and global security. In order to measure progress any technology provider, engineer or researcher working on biometrics will need to benchmark achievable system performance. Although there have been a number of challenges, competitions, and other efforts to achieve benchmarking in the past, the development of statistical tools, procedures and techniques for evaluating biometric performance has arguably not followed at the same pace. <br> <br>* How to evaluate and/or predict biometric system performance under covariates such as biometric sample quality, gender, age, etc? <br>* How to evaluate biometric performance under zero-effort and spoof attacks? <br>* How to evaluate the performance of biometric systems on a set with unreliable labels or without labels at all ? <br>* How to estimate and evaluate confidence intervals under varying samp... Read more

Special Issue: Biometric Performance and Statistics <br> <br>Biometrics technology increasingly plays a central role in personal, national and global security. In order to measure progress any technology provider, engineer or researcher working on biometrics will need to benchmark achievable system performance. Although there have been a number of challenges, competitions, and other efforts to achieve benchmarking in the past, the development of statistical tools, procedures and techniques for evaluating biometric performance has arguably not followed at the same pace. <br> <br>* How to evaluate and/or predict biometric system performance under covariates such as biometric sample quality, gender, age, etc? <br>* How to evaluate biometric performance under zero-effort and spoof attacks? <br>* How to evaluate the performance of biometric systems on a set with unreliable labels or without labels at all ? <br>* How to estimate and evaluate confidence intervals under varying sample population size? <br>* How does the biometric menagerie influence the system performance? <br> <br>This Special Issue will focus on the latest research on topics in biometric benchmarking and performance evaluation, with a particular emphasis on evaluation methodologies and statistics that render the performance metrics more generalizable, more accurate, and more reliable. The questions raised above are independent of any biometric modality. This field is closely tied to statistics, yet deeply rooted in biometrics. <br> <br>The principal aim of this Special Issue is to advance the state-of-the-art in this field, to stimulate interactions between the biometrics and the statistics communities, to encourage the development of reliable methodologies for performance assessment, and to promote the development of generalized performance metrics. <br> <br>Papers considered particularly relevant to this call are those that address biometric performance estimation, prediction, modelling, and methodology involving some of the following (non-exclusive) list of components: <br> <br>* Quality metrics, biometric sample quality, and utility <br>* Confidence estimation of performance metrics <br>* Prediction and performance extrapolation of unimodal and multimodal biometric systems n Biometric uniqueness and persistence <br>* Platforms, toolboxes, framework and methodology for evaluating biometric system performance <br>* Privacy and security assessment of biometric systems <br>* Case studies of performance assessments <br> <br>Contributions presented at recent relevant specialist conferences (for example, the International Biometrics Performance Conference, 2014) are welcomed, provided that they can demonstrate significant development and new material, as of course are entirely new and original contributions. <br> <br>All papers must be submitted through the journal’s Manuscript Central system: http://mc.manuscriptcentral.com/iet-bmt <br> <br>For enquiries regarding this Special Issue please contact Guest Editors: <br> <br>Norman Poh <br>University of Surrey, UK <br>E: normanpoh@ieee.org <br> <br>Michael Schuckers <br>St. Lawrence University <br>USA <br> <br>Honza Cernocky <br>Brno University <br>Czech Republic