Transformation, Ranking, and Clustering for Face Recognition Algorithm Performance

Published

January 1, 2002

Author(s)

Stefan D. Leigh, Nathanael A. Heckert, Andrew L. Rukhin, P J. Phillips, Patrick J. Grother, E M. Newton, M Moody, K Kniskern, S Heath

Abstract

The performance of face recognition algorithms is recently of increased interest, although to date empirical analyses of algorithms have been limited to rank-based scores such a cumulative match score and receiver operating characteristic. This paper demonstrates that algorithms that report ratio scale similarities between unknown and gallery images can be enormalized so that a large body of classical statistical methods can be applied to measure recognition performance.

Citation

Automatic Identification Advanced Technologies Workshop

Pub Type

Journals

Keywords

ANOVA, biometric evaluation, face recognition

Citation

Leigh, S. , Heckert, N. , Rukhin, A. , Phillips, P. , Grother, P. , Newton, E. , Moody, M. , Kniskern, K. and Heath, S. (2002), Transformation, Ranking, and Clustering for Face Recognition Algorithm Performance, Automatic Identification Advanced Technologies Workshop (Accessed April 25, 2024)