The 2nd Workshop on Demographic Variation in the Performance of Biometric Systems

Important Information

  • Paper submission deadline: December 20th, 2019
  • When: Sunday, March 1 (9:00 AM - 5:00 PM)
  • Where: Aspen, CO
  • Contact: info@mdtf.org

The main WACV 2020 conference will be held on March 2-5.

Overview

Automation of biometric recognition has spurred its burgeoning use in the public sector with both commercial and government applications. With increasing deployments of the technology it is important to consider whether it performs equitably for all users. There is significant public interest in this issue, with a large number of recently published popular media articles regarding the topic of “bias” in biometric systems, special interest groups advocating for complete bans on the use of some biometric technologies because of these supposed biases, and congressional hearings into government uses of these technologies.

Despite the public interest, existing government and academic work in this area is largely disjoint. This workshop will provide a common forum for applied and academic researchers to communicate their perspectives and exchange ideas regarding how to measure and mitigate any demographic effects in biometric performance. The expected outcomes are to increase awareness of demographic effects as they pertain to deployed systems, to help establish a framework for understanding their ultimate causes, and to develop mitigations.

Topics of Interest

The topics of interest for this workshop include the following:

    • Theory and nomenclature regarding demographic effects in biometric systems
    • Methods of measuring demographic factors and their utility in predicting biometric performance (e.g. phenotypes, social categories, self-reported vs. labeled)
    • Statistical modeling of performance variation in biometric systems due to demographic factors
    • Techniques for mitigating demographic effects on sensors and biometric recognition algorithms
    • Differences in demographic effects across different biometric modalities
    • Biological origins of demographic effects on physical and behavioral biometric traits
    • Automated extraction of demographic variables and soft biometric information from biometric data
    • Methods for fusing demographic information with biometric data
    • Imparting soft biometric privacy to biometric data

Call for Papers

This workshop will be the second of its series (first held at WACV 2019) to focus on the performance of biometric recognition systems across different demographic groups, while developing methods to detect and mitigate any biases. The organizers welcome high-quality paper submissions on the topics of interest (above). Selected papers will be provided the opportunity to present their findings to the workshop and have their work included in the published workshop proceedings. The full call for papers can be found here.

Format and Schedule

The 2nd Workshop on Demographic Variation in the Performance of Biometric Systems will be a full day workshop. The first portion of the day will feature presentations from the organizers and invited speakers, followed by a no-host lunch. The after-lunch session will include presentations from accepted papers, and a panel to discuss conclusions and future directions.

Workshop proceedings

The organizers will compile a list of notes from the workshop and make them publicly available, possibly in a special issue of IEEE Transactions on Biometrics, Behavior, and Identity Science (T-BIOM). Individual papers accepted for presentation at the workshop will be published as part of the WACV 2020 Workshop Proceedings and will be included in IEEE Xplore.

For information about WACV, please visit the main WACV 2020 Website.