This Article 
   
 Share 
   
 Bibliographic References 
   
 Add to: 
 
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
 
 Search 
   
Estimation of Nonlinear Errors-in-Variables Models for Computer Vision Applications
October 2006 (vol. 28 no. 10)
pp. 1537-1552
Bogdan C. Matei, IEEE Computer Society
Peter Meer, IEEE
In an errors-in-variables (EIV) model, all the measurements are corrupted by noise. The class of EIV models with constraints separable into the product of two nonlinear functions, one solely in the variables and one solely in the parameters, is general enough to represent most computer vision problems. We show that the estimation of such nonlinear EIV models can be reduced to iteratively estimating a linear model having point dependent, i.e., heteroscedastic, noise process. Particular cases of the proposed heteroscedastic errors-in-variables (HEIV) estimator are related to other techniques described in the vision literature: the Sampson method, renormalization, and the fundamental numerical scheme. In a wide variety of tasks, the HEIV estimator exhibits the same, or superior, performance as these techniques and has a weaker dependence on the quality of the initial solution than the Levenberg-Marquardt method, the standard approach toward estimating nonlinear models.

[1] K. Arun, T. Huang, and S. Blostein, “Least-Squares Fitting of Two 3D Point Sets,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 9, no. 5, pp. 698-700, Sept. 1987.
[2] J.W. Brewer, “Kronecker Products and Matrix Calculus in System Theory,” IEEE Trans. Circuits and Systems, vol. 25, no. 9, pp. 772-781, June 1978.
[3] J. Bride and P. Meer, “Registration via Direct Methods: A Statistical Approach,” Proc. 2001 IEEE Conf. Computer Vision and Pattern Recognition, pp. 984-989, Dec. 2001.
[4] H. Chen and P. Meer, “Robust Regression with Projection Based M-Estimators,” Proc. Ninth Int'l Conf. Computer Vision, vol. 2, pp. 878-885, Oct. 2003.
[5] W. Chojnacki, M.J. Brooks, and A. van den Hengel, “Rationalising the Renormalisation Method of Kanatani,” J. Math. Imaging and Vision, vol. 14, pp. 21-38, Feb. 2001.
[6] W. Chojnacki, M.J. Brooks, A. van den Hengel, and D. Gawley, “On the Fitting of Surfaces to Data with Covariances,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 22, no. 11, pp. 1294-1303, Nov. 2000.
[7] W. Chojnacki, M.J. Brooks, A. van den Hengel, and D. Gawley, “From FNS to HEIV: A Link between Two Vision Parameter Estimation Methods,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 26, no. 2, pp. 264-268, Feb. 2004.
[8] W. Chojnacki, M.J. Brooks, A. van den Hengel, and D. Gawley, “A New Constrained Parameter Estimator for Computer Vision Applications,” Image and Vision Computing, vol. 22, pp. 85-91, 2004.
[9] A. Edelman, T.A. Arias, and S.T. Smith, “The Geometry of Algorithms with Orthogonality Constraints,” SIAM J. Matrix Analysis Application, vol. 20, pp. 217-234, 1997.
[10] D. Eggert, A. Lorusso, and R. Fisher, “Estimating 3D Rigid Body Transformations: A Comparison of Four Major Algorithms,” Machine Vision and Applications, vol. 9, pp. 272-290, June 1997.
[11] W. Förstner, “On Estimating 3D Points and Lines from 2D Points and Lines,” Festschrift Anläßlich des 60. Geburtstages von Prof. Dr. Ing. Bernhard Wrobel, Technische Univ. Darmstadt, pp. 69-87, 2001.
[12] W. Fuller, Measurement Error Models. John Wiley and Sons, 1987.
[13] B. Georgescu and P. Meer, “Balanced Recovery of 3D Structure and Camera Motion from Uncalibrated Image Sequences,” Proc. European Conf. Computer Vision, vol. II, pp. 294-308, 2002.
[14] G. Golub and C.F. Van Loan, Matrix Computations, third ed. The Johns Hopkins Univ. Press, 1996.
[15] A. Graham, Kronecker Products and Matrix Calculus: With Applications. Ellis Horwood series in mathematics and its applications, 1981.
[16] R. Hartley and A. Zisserman, Multiple View Geometry in Computer Vision. Cambridge Univ. Press, 2000.
[17] B. Horn, H. Hilden, and S. Negahdaripour, “Closed-Form Solution of Absolute Orientation Using Orthonormal Matrices,” J. Optical Soc. of Am., vol. 5, pp. 1127-1135, 1988.
[18] S.V. Huffel and J. Vandewalle, “Analysis and Properties of GTLS in Problem $AX \approx B$ When Some or All Columns of A Are Subject to Errors,” SIAM J. Matrix Analysis and Applications, vol. 10, pp. 294-315, July 1989.
[19] K. Kanatani, Geometric Computation for Machine Vision. Oxford Science Publications, 1993.
[20] K. Kanatani, Statistical Optimization for Geometric Computation: Theory and Practice. Elsevier, 1996.
[21] K. Kanatani and D. Morris, “Gauges and Gauge Transformations for Uncertainty Description of Geometric Structure with Indeterminacy,” IEEE Trans. Information Theory, vol. 47, pp. 2017-2028, 2001.
[22] J. Koenderink and A. van Doorn, “The Generic Bilinear Calibration-Estimation Problem,” Int'l J. Computer Vision, vol. 23, no. 2, pp. 303-353, 1998.
[23] Y. Leedan and P. Meer, “Heteroscedastic Regression in Computer Vision: Problems with Bilinear Constraint,” Int'l J. Computer Vision, vol. 37, no. 2, pp. 127-150, June 2000.
[24] Y. Ma, J. Kosecka, and S. Sastry, “Optimization Criteria and Geometric Algorithms for Motion and Structure Estimation,” Int'l J. Computer Vision, vol. 44, pp. 219-249, 2001.
[25] B. Matei, “Heteroscedastic Errors-in-Variables Models in Computer Vision,” PhD thesis, Dept. of Electrical and Computer Eng., Rutgers Univ., 2001, http://www.caip.rutgers.edu/riul/research theses.html.
[26] B. Matei, B. Georgescu, and P. Meer, “A Versatile Method for Trifocal Tensor Estimation,” Proc. Eighth Int'l Conf. Computer Vision, vol. II, pp. 578-585, July 2001.
[27] B. Matei and P. Meer, “Optimal Rigid Motion Estimation and Performance Evaluation with Bootstrap,” Proc. Computer Vision and Pattern Recognition Conf., pp. 339-345, June 1999.
[28] B. Matei and P. Meer, “A General Method for Errors-in-Variables Problems in Computer Vision,” Proc. Computer Vision and Pattern Recognition Conf., vol. II, pp. 18-25, June 2000.
[29] B. Matei and P. Meer, “Reduction of Bias in Maximum Likelihood Ellipse Fitting,” Proc. 15th Int'l Conf. Computer Vision and Pattern Recognition, vol. III, pp. 802-806, Sept. 2000.
[30] B. Matei, P. Meer, and D. Tyler, “Performance Assessment by Resampling: Rigid Motion Estimators,” Empirical Evaluation Techniques in Computer Vision, pp. 72-95, 1998.
[31] D. Morris, K. Kanatani, and T. Kanade, “Gauge Fixing for Accurate 3D Estimation,” Proc. IEEE Conf. Computer Vision and Pattern Recognition, vol. II, pp. 343-350, 2001.
[32] O. Nestares and D. Fleet, “Errors-in-Variables Likelihood Functions for Motion Estimation,” Proc. IEEE Int'l Conf. Image Processing, 2003.
[33] N. Ohta and K. Kanatani, “Optimal Estimation of Three-Dimensional Rotation and Reliability Evaluation,” Proc. European Conf. Computer Vision (ECCV '98), pp. 175-187, June 1998.
[34] W. Press, S. Teukolsky, W. Vetterling, and B. Flannery, Numerical Recipes in C, second ed. Cambridge Univ. Press, 1992.
[35] P. Sampson, “Fitting Conic Sections to “Very Scattered” Data: An Iterative Refinement of the Bookstein Algorithm,” Computer Graphics and Image Processing, vol. 18, no. 1, pp. 97-108, Jan. 1982.
[36] G.W. Stewart, “Errors in Variables for Numerical Analysis,” Recent Advances in Total Least Squares Techniques and Errors-in-Variables Modeling, pp. 3-10, 1997.
[37] R. Subbarao, P. Meer, and Y. Genc, “A Balanced Approach to 3D Tracking from Image Streams,” Proc. IEEE and ACM Int'l Symp. Mixed and Augmented Reality, pp. 70-78, Oct. 2005.
[38] B. Triggs, P.F. McLauchlan, R.I. Hartley, and A.W. Fitzgibbon, “Bundle Adjustment— A Modern Synthesis,” Vision Algorithms: Theory and Practice, pp. 298-372, 2000.
[39] E. Trucco and A. Verri, Introductory Techniques for 3-D Computer Vision. Prentice Hall, 1998.
[40] S. Umeyama, “Least-Squares Estimation of Transformation Parameters between Two Point Patterns,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 13, no. 4, pp. 376-380, Apr. 1991.
[41] A. van den Hengel, W. Chonacki, M.J. Brooks, and D. Gawley, “A New Constrained Parameter Estimator: Experiments in Fundamental Matrix Computation,” Proc. British Machine Vision Conf., pp. 468-476, Sept. 2002.
[42] S. Van Huffel and J. Vandewalle, The Total Least Squares Problem. Computational Aspects and Analysis. Soc. for Industrial and Applied Math., 1991.
[43] R.H. Zamar, “Robust Estimation in the Errors-in-Variables Model,” Biometrika, vol. 76, pp. 149-160, 1989.
[44] Z. Zhang, “Image Matching Software,” 1995, http://www. inria.fr/robotvis/personnelzzhang .

Index Terms:
Nonlinear least squares, heteroscedastic regression, camera calibration, 3D rigid motion, uncalibrated vision.
Citation:
Bogdan C. Matei, Peter Meer, "Estimation of Nonlinear Errors-in-Variables Models for Computer Vision Applications," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 28, no. 10, pp. 1537-1552, Oct. 2006, doi:10.1109/TPAMI.2006.205
Usage of this product signifies your acceptance of the Terms of Use.