© The Institution of Engineering and Technology
The Foley-Sammon transform (FST) is one of the most well-known dimensionality reduction and feature extraction methods. However, the classical FST cannot be used directly in the small sample size problem where the within-class scatter matrix is singular. Null-space based FST (NFST) provides a good solution to this problem. Proposed is a fast incremental NFST (INFST). INFST extracts new information brought by newly-added samples and integrates it with the existing model by an efficient updating scheme. INFST can achieve the aims of online classification and novelty detection. Experiments on real-world data demonstrate the effectiveness of INFST.
References
-
-
1)
-
3. Bodesheim, P., Freytag, A., Rodner, E., et al: ‘Kernel null space methods for novelty detection’. CVPR 2013, Portland, OR, USA, June 2013, pp. 3374–3381.
-
2)
-
4. Pang, S., Ozawa, S., Kasabov, N.: ‘Incremental linear discriminant analysis for classification of data streams’, IEEE Trans. Syst. Man Cybern. B, Cybern., 2005, 35, (5), pp. 905–914 (doi: 10.1109/TSMCB.2005.847744).
-
3)
-
1. Guo, Y.-F., Wu, L., Lu, H., et al: ‘Null Foley Sammon transform’, Pattern Recogn., 2006, 39, (11), pp. 2248–2251 (doi: 10.1016/j.patcog.2006.05.009).
-
4)
-
5. Zhao, H., Yuen, P.C.: ‘Incremental linear discriminant analysis for face recognition’, IEEE Trans. Syst. Man Cybern. B, Cybern., 2008, 38, (1), pp. 210–221 (doi: 10.1109/TSMCB.2007.908870).
-
5)
-
2. Chu, D., Thye, G.S.: ‘A new and fast implementation for null space based linear discriminant analysis’, Pattern Recogn., 2010, 43, (4), pp. 1373–1379 (doi: 10.1016/j.patcog.2009.10.004).
-
6)
http://iet.metastore.ingenta.com/content/journals/10.1049/el.2015.2053
Related content
content/journals/10.1049/el.2015.2053
pub_keyword,iet_inspecKeyword,pub_concept
6
6