Author : Nitin Kumar 1
Date of Publication :22nd February 2018
Abstract: One of the problems in face recognition is the limited number of images per person available for training. In this paper, we investigate the performance of popular feature extraction methods such as Gabor wavelets, Discrete wavelet transform, Multi-view canonical correlation analysis, Linear discriminant analysis, Generalized uncorrelated linear discriminant analysis and Supervised canonical correlation analysis for face recognition with variation in the number of training images per person. The performance is measured in terms of classification accuracy. Experimental results on four publicly available datasets viz., AR, ORL, CMU-PIE and YALE demonstrate that the classification accuracy in general increases with increase in the number of training images per person with few exceptions.
Reference :
-
- W. Zhao, R. Chellappa, P.J. Philips, and A. Rosenfeld, "Face Recognition: A literature survey," ACM Computing Surveys, 35, 4, 399-458, 2003.
- S. Prabhakar, S. Pankanti, and A. K. Jain, "Biometric Recognition: Security and Privacy Concerns,” IEEE Security and Privacy, 1, 2, 33-42, 2003.
- A. Cordiner, “Illumination Invariant Face Detection,” MComSc Thesis, University of Wollongong, 2009.
- A. Pentland, and M. Turk, “Eigenfaces for Recognition,” Journal of Cognitive Neuroscience 3, 71- 86, 1991.
- P. Belhumeur, N. Peter, J. Hespanha, and D. J. Kriegman, "Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Linear Projection,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 19, 7, 711-720, 1997.