Abstract
A non-negative sparse semi-supervised dimensionality reduction framework is proposed for hyperspectral data. The framework consists of two parts: 1) a discriminant item is designed to analyze the few labeled samples from the global viewpoint, which can assess the separability between each surface object; 2) a regularization term is used to build a non-negative sparse representation graph based on large scale unlabelled samples, which can adaptively find an adjacency graph for each sample and then find valuable samples from the original hyperspectral data. Based on the framework and the maximum margin criterion, a dimensionality reduction algorithm called non-negative sparse semi-supervised maximum margin criterion is proposed. Experimental results on the AVIRIS 92AV3C hyperspectral data show that the proposed algorithm can effectively utilize the unlabelled samples to obtain higher overall classification accuracy.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Cai, D., He, X.F., Han, J.W.: Semi-supervised discriminant analysis. In: Proceedings of the 11th IEEE International Conference on Computer Vision, pp. 222–228. IEEE, Brazil (2007)
Song, Y.Q., Nie, F.P., Zhang, C.S.: A Unified framework for semi-supervised dimensionality reduction. Pattern Recognition 41(9), 2789–2799 (2008)
Farcas, D., Marghes, C., Bouwmans, T.: Background subtraction via incremental maximum margin criterion: a discriminative subspace approach. Machine Vision and Applications 23(6), 1083–1101 (2012)
Zhu, X.: Semi-supervised learning literature survey. University of Wisconsin, Madison (2008)
Qiao, L.S., Chen, S.C., Tan, X.Y.: Sparsity Preserving discriminant analysis for single training image face recognition. Pattern Recognition Letters 31(5), 422–429 (2010)
Gui, J., Sun, Z.N., Jia, W.: Discriminant sparse neighborhood preserving embedding for face recognition. Pattern Recognition 45(8), 2884–2893 (2012)
Wong, W.K.: Discover latent discriminant information for dimensionality reduction: non-negative sparseness preserving embedding. Pattern Recognition 45(4), 1511–1523 (2012)
Lv, J.Q., Pawlak, M., Annakkage, U.D.: Prediction of the transient stability boundary using the lasso. IEEE Transactions on Power Systems 28(1), 281–288 (2013)
Zou, H., Hastie, T.: Regularization and variable selection via the elastic Net. Journal of the Royal Statistical Society Series B-Statistical Methodology 67(2), 301–320 (2005)
He, R., Hu, B.G., Zheng, W.S.: Two-stage sparse representation for robust recognition on large-scale database. In: Proceedings of the 24th AAAI Conference on Artificial Intelligence, pp. 475–480. American Association for Artificial Intelligence, Atlanta (2010)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer International Publishing Switzerland
About this paper
Cite this paper
Wang, X., Gao, Y., Cheng, Y. (2014). Hyperspectral Data Dimensionality Reduction Based on Non-negative Sparse Semi-supervised Framework. In: Huang, DS., Bevilacqua, V., Premaratne, P. (eds) Intelligent Computing Theory. ICIC 2014. Lecture Notes in Computer Science, vol 8588. Springer, Cham. https://doi.org/10.1007/978-3-319-09333-8_86
Download citation
DOI: https://doi.org/10.1007/978-3-319-09333-8_86
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-09332-1
Online ISBN: 978-3-319-09333-8
eBook Packages: Computer ScienceComputer Science (R0)