CONSTRAINED GENERALISED PRINCIPAL COMPONENT ANALYSIS

Wojciech Chojnacki, Anton van den Hengel, Michael J. Brooks

Abstract

Generalised Principal Component Analysis (GPCA) is a recently devised technique for fitting a multi-component, piecewise-linear structure to data that has found strong utility in computer vision. Unlike other methods which intertwine the processes of estimating structure components and segmenting data points into clusters associated with putative components, GPCA estimates a multi-component structure with no recourse to data clustering. The standard GPCA algorithm searches for an estimate by minimising an appropriate misfit function. The underlying constraints on the model parameters are ignored. Here we promote a variant of GPCA that incorporates the parameter constraints and exploits constrained rather than unconstrained minimisation of the error function. The output of any GPCA algorithm hardly ever perfectly satisfies the parameter constraints. Our new version of GPCA greatly facilitates the final correction of the algorithm output to satisfy perfectly the constraints, making this step less prone to error in the presence of noise. The method is applied to the example problem of fitting a pair of lines to noisy image points, but has potential for use in more general multi-component structure fitting in computer vision.

References

  1. (2003). Proc. IEEE Conf. Computer Vision and Pattern Recognition.
  2. Brooks, M. J., Chojnacki, W., Gawley, D., and van den Hengel, A. (2001). What value covariance information in estimating vision parameters? In Proc. Eighth Int. Conf. Computer Vision, volume 1, pages 302-308.
  3. Chojnacki, W., Brooks, M. J., van den Hengel, A., and Gawley, D. (2000). On the fitting of surfaces to data with covariances. IEEE Trans. Pattern Anal. Mach. Intell., 22(11):1294-1303.
  4. Chojnacki, W., Brooks, M. J., van den Hengel, A., and Gawley, D. (2004a). From FNS and HEIV: A link between two vision parameter estimation methods. IEEE Trans. Pattern Anal. Mach. Intell., 26(2):264-268.
  5. Chojnacki, W., Brooks, M. J., van den Hengel, A., and Gawley, D. (2004b). A new constrained parameter estimator for computer vision applications. Image and Vision Computing, 22:85-91.
  6. Chojnacki, W., Brooks, M. J., van den Hengel, A., and Gawley, D. (2005). FNS, CFNS and HEIV: A unifying approach. J. Math. Imaging and Vision, 23(2):175-183.
  7. Duda, R. O. and Hart, P. E. (1972). Use of the Hough transform to detect lines and curves in pictures. Commun. ACM, 15:11-15.
  8. Fitzgibbon, A., Pilu, M., and Fisher, R. B. (1999). Direct least square fitting of ellipses. IEEE Trans. Pattern Anal. Mach. Intell., 21(5):476-480.
  9. Forsyth, D. A. and Ponce, J. (2003). Computer Vision: A Modern Approach. Prentice Hall.
  10. Halír?, R. and Flusser, J. (1998). Numerically stable direct least squares fitting of ellipses. In Proc. Sixth Int. Conf. in Central Europe on Computer Graphics and Visualization, pages 125-132.
  11. Ho, J., Yang, M.-H., Lim, J., Lee, K.-C., and Kriegman, D. J. (2003). Clustering appearances of objects under varying illumination conditions. In (CVP, 2003), pages 11-18.
  12. Horn, R. and Johnson, C. (1985). Matrix Analysis. Cambridge University Press, Cambridge.
  13. Kanatani, K. (1996). Statistical Optimization for Geometric Computation: Theory and Practice. Elsevier, Amsterdam.
  14. Leedan, Y. and Meer, P. (2000). Heteroscedastic regression in computer vision: Problems with bilinear constraint. Int. J. Computer Vision, 37(2):127-150.
  15. Leonardis, A., Bischof, H., and Maver, J. (2002). Multiple eigenspaces. Pattern Recognition, 35(11):2613-2627.
  16. Lim, M. H. (1992). Conditions on decomposable symmetric tensors as an algebraic variety. Linear and Multilinear Algebra, 32:249-252.
  17. Lou, X.-M., Hassebrook, L. G., Lhamon, M. E., and Li, J. (1997). Numerically efficient angle, width, offset, and discontinuity determination of straight lines by the discrete Fourier-bilinear transformation algorithm. IEEE Trans. Image Processing, 6(10):1464-1467.
  18. Matei, B. and Meer, P. (2000). A general method for errorsin-variables problems in computer vision. In Proc. IEEE Conf. Computer Vision and Pattern Recognition, volume 2, pages 18-25.
  19. Nievergelt, Y. (2004). Fitting conics of specific types to data. Linear Algebra and Appl., 378:1-30.
  20. O'Leary, P. and Zsombor-Murray, P. (2004). Direct and specific least-square fitting of hyperbolae and ellipses. J. Electronic Imaging, 13(3):492-503.
  21. Tipping, M. E. and Bishop, C. M. (1999). Mixtures of probabilistic principal component analysers. Neural Computation, 11(2):443-482.
  22. Venkateswar, V. and Chellappa, R. (1992). Extraction of straight lines in aerial images. IEEE Trans. Pattern Anal. Mach. Intell., 14(11):1111-1114.
  23. Vidal, R. and Ma, Y. (2004). A unified algebraic approach to 2-D and 3-D motion segmentation. In Proc. Eighth European Conf. Computer Vision, volume 3021 of Lecture Notes on Computer Vision, pages 1-15.
  24. Vidal, R., Ma, Y., and Piazzi, J. (2004). A new GPCA algorithm for clustering subspaces by fitting, differentiating and dividing polynomials. In Proc. IEEE Conf. Computer Vision and Pattern Recognition, volume 1, pages 510-517.
  25. Vidal, R., Ma, Y., and Sastry, S. (2003). Generalized principal component analysis (GPCA). In (CVP, 2003), pages 621-628.
  26. Vidal, R., Ma, Y., and Sastry, S. (2005). Generalized principal component analysis (GPCA). IEEE Trans. Pattern Anal. Mach. Intell., 27(12):1945-1959.
  27. Vidal, R., Ma, Y., Soatto, S., and Sastry, S. (2002). Segmentation of dynamical scenes from the multibody fundamental matrix. In ECCV Workshop “Vision and Modelling of Dynamical Scenes”. Available at http://www.robots.ox.ac.uk/ ~awf/eccv02/vamods02-rvidal.pdf.
  28. Vidal, R., Ma, Y., Soatto, S., and Sastry, S. (2006). Twoview multibody structure from motion. Int. J. Computer Vision. In press.
Download


Paper Citation


in Harvard Style

Chojnacki W., van den Hengel A. and J. Brooks M. (2006). CONSTRAINED GENERALISED PRINCIPAL COMPONENT ANALYSIS . In Proceedings of the First International Conference on Computer Vision Theory and Applications - Volume 1: VISAPP, ISBN 972-8865-40-6, pages 206-212. DOI: 10.5220/0001362102060212


in Bibtex Style

@conference{visapp06,
author={Wojciech Chojnacki and Anton van den Hengel and Michael J. Brooks},
title={CONSTRAINED GENERALISED PRINCIPAL COMPONENT ANALYSIS},
booktitle={Proceedings of the First International Conference on Computer Vision Theory and Applications - Volume 1: VISAPP,},
year={2006},
pages={206-212},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0001362102060212},
isbn={972-8865-40-6},
}


in EndNote Style

TY - CONF
JO - Proceedings of the First International Conference on Computer Vision Theory and Applications - Volume 1: VISAPP,
TI - CONSTRAINED GENERALISED PRINCIPAL COMPONENT ANALYSIS
SN - 972-8865-40-6
AU - Chojnacki W.
AU - van den Hengel A.
AU - J. Brooks M.
PY - 2006
SP - 206
EP - 212
DO - 10.5220/0001362102060212