SLU Statistics Professor Presents Work on Testing Equity in Biometric Recognition Systems

Professor Michael Schuckers presented a paper at a recent workshop on bias in biometric authentication systems.  The workshop entitled “Understanding and Mitigating Demographic Bias in Biometric Systems” (https://vcbsl-wsu.github.io/icpr22w-umdbb/) was part of the 2022 International Conference on Pattern Recognition (https://www.icpr2022.com/) which was recently held in Montréal, Québec. 

Biometric systems include technology such as fingerprint readers or facial recognition algorithms.   

 Schuckers paper focused on methods for determining if there are statistical differences between demographic groups for these systems on their false reject rates.  Schuckers was the lead author and was joined by colleagues at Clarkson University and the Center for Identification Technology Research (CITeR).  His co-authors are Sandip Purnapatra, Kaniz Fatim, Daqing Hou and Stephanie Schuckers. 

Recently, a good deal of attention has been paid to differences in performance of biometrics technology between racial and ethnic groups, especially regarding facial recognition systems.  In their paper, Schuckers and collaborators presented some of the first methodologies for statistically testing these differences.   

Ensuring that false reject rates are consistent across groups is important for making these tools fairly and evenly accessible.    

This work was funded in part by grants from the CITeR and the National Science Foundation. 

Schuckers is Charles A Dana Professor of Statistics and department chair for Mathematics, Computer Science and Statistics at St. Lawrence University.  He is also author of the 2010 book Computational Methods in Biometric Authentication: Statistical Methods for Performance Evaluation which is foundational to the assessment of biometric authentication systems.