Ongoing Face Recognition Vendor Test (FRVT) part 6A:
Mei Lee Ngan, Patrick J. Grother, Kayee Hanaoka · 2020
This is the first of a series of reports on the performance of face recognition algorithms on faces occluded by protective face masks [2] commonly worn to reduce inhalation of viruses or other contaminants.This study is being run under the Ongoing Face Recognition Vendor Test (FRVT) executed by the National Institute of Standards and Technology (NIST).This report documents accuracy of algorithms to recognize persons wearing face masks.The results in this report apply to algorithms provided to NIST before the COVID-19 pandemic, which were developed without expectation that NIST would execute them on masked face images.NIST had informed the FRVT developer community of our intent to run existing algorithms on masked images prior to the outset of this study and invited submission of mask-enabled algorithms for the next phase of this work.This report is intended to support end-users to understand how a pre-pandemic algorithm might be affected by the arrival of a substantial number of subjects wearing face masks.The next report will document accuracy values for more recent algorithms, some developed with capabilities for recognition of masked faces.The algorithms tested were one-to-one algorithms submitted to the FRVT 1:1 Verification track.Future iterations of this document will also report accuracy of one-to-many algorithms.