Visual Analysis of the Use of Mixture Covariance Matrices in Face Recognition
The quadratic discriminant (QD) classifier has proved to be simple and effective in many pattern recognition problems. However, it requires the computation of the inverse of the sample group covariance matrix. In many biometric problems, such as face recognition, the number of training patterns is considerably smaller than the number of features, and therefore the covariance matrix is singular. Several studies have shown that the use of mixture covariance matrices defined as a combination between the sample group covariance matrices and, for instance, the pooled covariance matrix, not only overcomes the singularity and instability of the sample group covariance matrices but also improves the QD classifier performance. However, little attention has been paid to understanding what has happened with the final shape of these mixture covariance matrices. In this work, we visually analyse in the commonly used eigenfaces space the eigenvectors and eigenvalues of these covariance matrices, given by the three following approaches: maximum likelihood, maximum classification accuracy, and maximum entropy. Experiments using the two well-known ORL and FERET face databases have shown that the maximum entropy approach is the one that achieves a more intuitive visual performance and best classification accuracies, especially in face experiments where moderate changes in facial expressions, pose, and scale, occur.
KeywordsFace Recognition Covariance Matrice Good Classification Accuracy Mixture Parameter FERET Database
Unable to display preview. Download preview PDF.
- C. Liu and H. Wechsler, “Probabilistic Reasoning Models for Face Recognition”. In Proc. of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR’98, Santa Barbara, California, USA, June 1998.Google Scholar
- C. E. Thomaz, D. F. Gillies and R. Q. Feitosa, “Using Mixture Covariance Matrices to Improve Face and Facial Expression Recognitions”, in Proc. of 3rd International Conference of Audio-and Video-Based Biometric Person Authentication AVBPA’01, Springer-Verlag LNCS 2091, pp. 71–77, Halmstad, Sweden, June 2001.Google Scholar
- C. E. Thomaz, D. F. Gillies and R. Q. Feitosa. A New Quadratic Classifier applied to Biometric Recognition. In proceedings of the Post-ECCV Workshop on Biometric Authentication, Springer-Verlag LNCS 2359, pp. 186–196, Copenhagen, Denmark, June 2002.Google Scholar
- E. T. Jaynes, “On the rationale of maximum-entropy methods”, Proceedings of the IEEE, vol. 70, pp. 939–952, 1982.Google Scholar
- J.P. Hoffbeck and D.A. Landgrebe, “Covariance Matrix Estimation and Classification with Limited Training Data”IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 18, no. 7, July 1996.Google Scholar
- J.R. Magnus and H. Neudecker, Matrix Differential Calculus with Applications in Statistics and Econometrics, revised edition. Chichester: John Wiley & Sons Ltd., 1999.Google Scholar
- K. Fukunaga, Introduction to Statistical Pattern Recognition, second edition. Boston: Academic Press, 1990.Google Scholar
- M. Kirby and L. Sirovich, “Application of the Karhunen-Loeve procedure for the characterization of human faces”, IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 12, No. 1, Jan. 1990.Google Scholar
- R.A. Johnson R.A. and D.W. Wichern, Applied Multivariate Statistical Analysis, by Prentice-Hall, Inc., 3d. edition, 1992.Google Scholar
- T.W. Anderson, An Introduction to Multivariate Statistical Analysis, second edition. New York: John Wiley & Sons, 1984.Google Scholar