Abstract
In this paper, we will consider a robust estimator, which was proposed earlier by the authors, in a general non-linear regression framework. The basic idea of the estimator is, instead of trying to classify the observations to good and false, to model the residual distribution of the contaminants, determine the probability for each observation to be a good sample, and finally perform weighted fitting. The main contributions of this paper are: (1) We show now that the estimator is consistent with the true parameter values that simply means optimality regardless of the problematical outliers in the observations. (2) We propose how robust uncertainty computations and robust model selection can be performed in the similar, consistent manner. (3) We derive the expectation maximisation algorithm for the estimator and (4) extend the estimator to handle unknown outlier residual distributions. (5) We finally give some experiments with real data, where robustness in model fitting is needed.
Similar content being viewed by others
References
H. Akaike, “On entropy maximization principle,” in (ed.), Applications of Statistics, P.R. Krishnaiah, North Holland, Amsterdam, pp. 27–41, 1977.
R. Beran, “Asymptotically efficient adaptive rank estimates in location models,” Ann. Stat. Vol. 2, No. 1, pp. 63–74, 1974.
C. Biernacki, G. Celeux, and G. Govaert, “Strategies for getting the largest likelihood in mixture models,” Presented at the invited session in Joint Statistical Meetings 2000, Indianapolis. Presentation available at http://www.inrialpes.fr/is2/people/celeux/, 2000.
S. Brandt, “Maximum likelihood robust regression with known and unknown residual models,” in Proc of the Statistical Methods in Video Proc Workshop, in conjunction with ECCV 2002, Copenhagen, Denmark, 2002a, pp. 97–102.
S. Brandt, “Theorems and algorithms for multiple view geometry with applications to electron tomography,” Doctoral thesis, Helsinki University of Technology, 2002b.
S. Brandt and J. Heikkonen, “A Bayesian weighting principle for the fundamental matrix estimation,” Pattern Recognit. Lett. Vol. 21, No. 12, pp. 1081–1092, 2000.
S. Brandt and J. Heikkonen, “Optimal method for the affine F-matrix and its uncertainty estimation in the sense of both noise and outliers,” in Proc. of the Eighth International Conference on Computer Vision, Vol. 2, Vancouver, Canada, 2001, pp. 166–173.
G. Celeux and J. Diebolt, “The SEM algorithm: A probabilistic teacher algorithm derived from the EM algorithm for the mixture problem,” Computat. Stat. Quart, Vol. 2, pp. 73–82, 1985.
G. Celeux and G. Govaert, “A classification EM algorithm for clustering and two stochastic versions,” Computat. Stat. Data Anal, Vol. 14, pp. 315–332, 1992
G. Csurka, C. Zeller, Z. Zhang, and O.D. Faugeras, “Characterizing the uncertainty of the fundamental matrix,” Comput. Vis. Image Underst, Vol. 68, No. 1, pp. 18–36, 1997
A.P. Dempster, N.M. Laird, and D.B. Rubin, “Maximum likelihood form incomplete data via the EM algorithm (with discussion),” J.R. Statist. Soc. Ser. B-Stat. Methodol, Vol. 39, pp. 1–38, 1977.
M. Fischler and L. Bolles, “Random sample consensus: A paradigm for model fitting with applications to image analysis and automated cartography,” Commun. ACM, Vol. 24, pp. 381–385, 1981.
C. Fraley and A.E. Raftery, “How many clusters? which clustering method? answers via model-based cluster analysis,” Comput. J, Vol. 41, pp. 578–588, 1998.
G.H. Golub and C.F.V. Loan, Matrix Computations, Second edn. The John Hopkins University Press, Chap. 7, 1989, pp. 345–346.
R. Hartley and A. Zisserman, Multiple View Geometry in Computer Vision, Cambridge University Press, 2000.
R.V. Hogg, “Adaptive robust procedures: A partial review and some suggestions for future applications and theory,” J. Am. Stat. Assoc, Vol. 69, No. 348, pp. 909–927, 1974.
P.J. Huber, Robust Statistics, Wiley, 1981.
INRIA Syntim Project, http://www-syntim.inria.fr/syntim/analyse/paires-eng.html. 1999.
H. Longuet-Higgins, “A computer algorithm for reconstructing a scene from two projections,” Nature, Vol. 293, pp. 133–135, 1981.
Q.-T. Luong and O. Faugeras, “The fundamental matrix: Theory, algorithms, and stability analysis,” Int. J. Comput. Vis., Vol. 17, No. 1, pp. 43–76, 1996.
G. McLachlan and T. Krishnan, The EM Algorithm and Extensions, Wiley, 1996.
J. Rissanen, “Modeling by shortest data description,” Automatica, Vol. 14, pp. 465–471, 1978.
J. Rissanen, “A universal prior for integers and estimation by minimum description length,” Ann. Stat., Vol. 11, No. 2, pp. 416–431, 1983.
J. Rissanen, Stochastic Complexity in Statistical Inquiry, Vol. 15 of Series in Computer Science, World Scientific, Singapore, 1989.
P.J. Rousseeuw, “Least median of squares regression,” J. Am. Stat. Assoc., Vol. 79, pp. 871–880, 1984.
P.J. Rousseeuw and A.M. Leroy, Robust Regression and Outlier Detection, Wiley, 1987.
J. Sacks, “An asymptotically efficient sequence of estimators of a location parameter,” Ann. Stat., Vol. 3, No. 2, pp. 285–298, 1975.
Y. Sakamoto, M. Ishiguro, and G. Kitagawa, Akaike Information Criterion Statistics, KTK Scientific Publishers, 1986.
G. Schwarz, “Estimating the dimension of a model,” Ann. Stat., Vol. 6, No. 2, pp. 461–464, 1978.
L. Shapiro and J. Brady, “Rejecting outliers and estimating errors in an orthogonal regression framework,” Philos. Trans. R. Soc. Lond. Ser. A-Math. Phys. Eng. Sci., Vol. 350, 1995.
L. Shapiro, A. Zisserman, and M. Brady, “Motion from point matches using affine epipolar geometry,” in Proceedings of the Third European Conference on Computer Vision, 1994.
M. Srinath, P. Rajasekaran, and R. Viswanathan, Introduction to Statistical Signal Processing with Applications, Prentice Hall, Chap. 5, 1996, pp. 146–149.
C.J. Stone, “Adaptive maximum likelihood estimators,” Ann. Stat., Vol. 3, No. 2, pp. 267–284, 1975.
P. Torr and D. Murray, “The development and comparison of robust methods for estimating the fundamental matrix,” Int. J. Comput. Vis., Vol. 24, No. 3, pp. 271–300, 1997.
P. Torr and A. Zisserman, “MLESAC: A new robust estimator with application to estimating image geometry,” Comput. Vis. Image Underst, Vol. 78, No. 1, pp. 138–156, 2000.
V.N. Vapnik, The Nature of Statistical Learning Theory, Second edn., Springer-Verlag, New York, 2000.
J. Weng, T.S. Huang, and N. Ahuja, “Motion and structure from two perspective views: Algorithms, error analysis, and error estimation,” IEEE Trans. Pattern Anal. Mach. Intell., Vol. 11, No. 5, pp. 451–476, 1989.
C.J. Wu, “On the convergence properties of the EM algorithm,” Ann. Stat, Vol. 11, No. 1, pp. 95–103, 1983.
G. Xu and Z. Zhang, Epipolar Geometry in Stereo, Motion and Object Recognition, Vol. 6 of Computational Imaging and Vision, Kluwer Academic Publishers, 1996.
Z. Zhang, R. Deriche, O. Faugeras, and Q. Luong, “A robust technique for matching two uncalibrated images through the recovery of the unknown epipolar geometry,” Artif. Intell., Vol. 78, pp. 87–119, 1994.
Z. Zhang and C. Loop, “Estimating the fundamental matrix by transforming image points in protective space,” Comput. Vis. Image Underst, Vol. 82, pp. 174–180, 2001.
Author information
Authors and Affiliations
Corresponding author
Additional information
Sami Brandt received the degree of Master of Science in Technology from the department of Engineering Physics and Mathematics in Helsinki University of Technology, Finland, in September 1999 and the degree of Doctor of Science in Technology at the Laboratory of Computational Engineering, Helsinki University of Technology, in October 2002. After serving one year as a research scientist in Instrumentarium Corporation Imaging Division and two years as a post-doc at LCE, he is currently jointly affiliated at LCE and Information Processing Laboratory, University of Oulu, Finland; and he focuses research on bio-medical imaging and 3D vision. He is a member of the IEEE and IEEE Computer Society, member of the Pattern Recognition Society of Finland, member of the International Association for Pattern Recognition (IAPR), and member of the Finnish Inverse Problems Society.
Rights and permissions
About this article
Cite this article
Brandt, S.S. Maximum Likelihood Robust Regression by Mixture Models. J Math Imaging Vis 25, 25–48 (2006). https://doi.org/10.1007/s10851-005-4386-4
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10851-005-4386-4