Communications of the ACM
NIST Evaluates Face Recognition Software's Accuracy for Flight Boarding
According to recent tests at the National Institute of Standards and Technology, face recognition algorithms have demonstrated the capability to confirm airline passenger identities while making very few errors.
Credit: N. Hanacek/NIST
Tests of face recognition (FR) software at the U.S. National Institute of Standards and Technology (NIST) indicate the most accurate algorithms are able to confirm airline passenger identities at boarding while committing few errors.
The NIST researchers assessed the algorithms' performance in matching images of travelers to previously acquired photos stored in a database.
They found the seven top-performing FR algorithms could find a match based on a single scan of a passenger's face with 99.5% or higher accuracy, particularly if several passenger images are in the database.
Explained NIST’s Patrick Grother, “We found that accuracy varies across algorithms, but that modern algorithms generally perform better. If airlines use the more accurate ones, passengers can board many flights with no errors.”
From U.S. National Institute of Standards and Technology
View Full Article
Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA
No entries found