Abstract
We present a novel detection approach, detection with canonical correlation (DCC), for target detection without prior information on the interference. We use the maximum canonical correlations between the target set and the observation data set as the detection statistic, and the coefficients of the canonical vector are used to determine the indices of components from a given target library, thus enabling both detection and classification of the target components that might be present in the mixture. We derive an approximate distribution of the maximum canonical correlation when targets are present. For applications where the contributions of components are non-negative, non-negativity constraints are incorporated into the canonical correlation analysis framework and a recursive algorithm is derived to obtain the solution. We demonstrate the effectiveness of DCC and its nonnegative variant by applying them on detection of surface-deposited chemical agents in Raman spectroscopy.
Similar content being viewed by others
Notes
This paper is an extension of our earlier work presented at the IEEE Workshop on Machine Learning for Signal Processing [11], where the idea of detection with canonical correlation for unknown interference was first introduced.
References
Kay, S. M. (1998). Fundamentals of statistical signal processing: Detection theory. Prentice Hall PTR, NJ.
Manolakis, D., Marden, D., & Shaw, G. A. (2003). Hyperspectral image processing for automatic target detection. MIT Lincoln Lab Journal, 14(1), 79–116.
ITT Industries (2003). Tests of the laser interrogation of surface agents system for on-the-move standoff sensing of chemical agents. In Proc. int. symp. spectral sensing research.
Scharf, L. L., & Friedlander, B. (1994). Matched subspace detectors. IEEE Transactions on Signal Processing, 42(8), 2146–2157.
Manolakis, D., et al. (2001). Hyperspectral subpixel target detection using the linear mixing model. IEEE Transactions on Geoscience and Remote Sensing, 39(7), 1392–1409.
Kraut, S., Scharf, L. L., & McWhorter, L. T. (2001). Adaptive subspace detectors. IEEE Transactions on Signal Processing, 49(1), 1–16.
Wang, W., & Adalı, T. (2007). Detection using correlation bound in a linear mixture model. Signal Processing, 87(5), 1118–1127.
Wang, W., & Adalı, T. (2005). Constrained ICA and its application to Raman spectroscopy. In Proc. antennas and propagation society international symposium (pp. 109–112). Washington, DC.
Li, H., Adalı, T., Wang, W., & Emge, D. (2007). Non-negative matrix factorization with orthogonality constraints and its application to Raman spectroscopy. Journal of VLSI Signal Processing, 48, 83–97.
Desai, M. N., & Mangoubi, R. S. (2003). Robust Gaussian and non-Gaussian matched subspace detection. IEEE Transactions on Signal Processing, 51(12).
Wang, W., Adalı, T., & Emge, D. (2007). Unsupervised detection using canonical correlation analysis and its application to Raman spectroscopy. In Proc. IEEE workshop on machine learning for signal processing, Thessaloniki, Greece.
Hotelling, H. (1936). Relations between two sets of variates. Biometrika, 28, 321–377.
Anderson, T. W. (2003). An introduction to multivariate statistical analysis. CA: Wiley.
Constantine, A. G. (1963). Some non-central distribution problems in multivariate analysis. Annals of Mathematical Statistics, 34(4), 1270–1285.
Hayakawa, T. (1967). On the distribution of the maximum latent root of a positive definite symmetric random matrix. Annals of the Institute of Statistical Mathematics, 21(1), 1–17.
Fisher, R. A. (1928). The general sampling distribution of the multiple correlation coefficient. Proceedings of Royal Society, A, 121, 654–673.
Tenenhaus, M. (1988). Canonical analysis of two convex polyhedral cones and applications. Psychometrika, 53(4), 503–524.
Vía, J., Santamaría, I., & Pérez, J. (2005). A robust RLS algorithm for adaptive canonical correlation analysis. In Proc. IEEE int. conf. acoust., speech, signal processing, 4, 365–368. Philadelphia, PA.
Kennedy, W. J., & Gentle, J. E. (1980). Statistical computing. New York: Marcel Dekker.
Slamani, M., Chyba, T., LaValley, H., & Emge, D. (2006). Identification algorithm for the joint contaminated surface detector (JCSD). In Proc. 2006 international symposium on spectral sensing research, Bar Harbor, ME.
Wang, W., Adalı, T., & Emge, D. (2009). Subspace partitioning for target detection and identification. IEEE Transactions on Signal Processing, 57(4), 1250–1259.
Wang, W., & Adalı, T. (2008). Target detection and identification using canonical correlation and subspace partitioning. In Proc. IEEE int. conf. acoust., speech, signal processing (pp. 2117–2120). Las Vegas, NV.
Author information
Authors and Affiliations
Corresponding author
Additional information
This project is supported by Edgewood Chemical and Biological Center, US Army RDECOM under contract no: W91ZLK-04-P-0950.
Appendix: Distribution of the Multiple Correlation
Appendix: Distribution of the Multiple Correlation
In this appendix, we evaluate the distribution of the square of multiple correlation, which can be written as [13]:
where \(\mathbf{y}\sim\mathcal{N}\left(\mathbf{t}_1,\sigma^2\mathbf{I}\right)\) and T is an any given matrix, such that T = [t 1, ⋯ ,t L ].
We rotate the coordinates of y so that the last L coordinates are in the subspace spanned by the columns of T. Given the singular value decomposition of \(\mathbf{T}^T\mathbf{T}=\mathbf{Q}{\boldsymbol{\Lambda}}\mathbf{Q}^T\), we define \(\mathbf{F}=\mathbf{Q}{\boldsymbol\Lambda}^{-\frac{1}{2}}\) so that H − 1 = FF T. We construct an N × N orthonormal matrix P = [P 1 P 2], where P 2 = TF, and u = P T y. We can see that
where \(\mathbf{e}_1=[1, 0, \cdots, 0]^T\), and \(\mathbf{f} \triangleq \mathbf{F}^{-1}\mathbf{e}_1\). Since P is orthonormal, the covariance matrix of u is the same with y, i.e., σ 2 I, thus, we have
where
Let
then
Considering Eqs. 17, 18 and u T u = y T PP T y = y T y, we can write
The numerator in Eq. 19 follows a non-central χ 2-distribution with L degrees of freedom, and the denominator follows a χ 2-distribution with N − L degrees of freedom. The numerator and denominator are independent since u i ’s are independently distributed.
We define
which has a non-central F-distribution with L and N − L degrees of freedom and a non-centrality parameter given by
The probability density function of h is
where \(\beta(x,y)=\int_0^1{t^{x-1}(1-t)^{y-1}dt}\) is the beta function. Since
then
The expectation of α can be obtained as:
The last step is obtained by using the following equalities:
where \(\Gamma(x)=\int_0^{\infty}{t^{x-1}e^{-t}dt}\) is the gamma function.
Rights and permissions
About this article
Cite this article
Wang, W., Adalı, T. & Emge, D. A Novel Approach for Target Detection and Classification Using Canonical Correlation Analysis. J Sign Process Syst 68, 379–390 (2012). https://doi.org/10.1007/s11265-011-0625-7
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11265-011-0625-7