Advertisement

Computing

, 86:73 | Cite as

Distance regression by Gauss–Newton-type methods and iteratively re-weighted least-squares

  • Martin Aigner
  • Bert Jüttler
Article

Abstract

We discuss the problem of fitting a curve or surface to given measurement data. In many situations, the usual least-squares approach (minimization of the sum of squared norms of residual vectors) is not suitable, as it implicitly assumes a Gaussian distribution of the measurement errors. In those cases, it is more appropriate to minimize other functions (which we will call norm-like functions) of the residual vectors. This is well understood in the case of scalar residuals, where the technique of iteratively re-weighted least-squares, which originated in statistics (Huber in Robust statistics, 1981) is known to be a Gauss–Newton-type method for minimizing a sum of norm-like functions of the residuals. We extend this result to the case of vector-valued residuals. It is shown that simply treating the norms of the vector-valued residuals as scalar ones does not work. In order to illustrate the difference we provide a geometric interpretation of the iterative minimization procedures as evolution processes.

Keywords

Curve and surface fitting Iteratively re-weighted least squares Gauss–Newton method Fitting by evolution 

Mathematics Subject Classification (2000)

65D17 68U07 

References

  1. 1.
    Aigner M, Jüttler B (2007) Approximation flows in shape manifolds. In: Chenin P, Lyche T, Schumaker L (eds) Curve and surface design: Avignon 2006, Nashboro Press, pp 1–10Google Scholar
  2. 2.
    Aigner M, Šír Z, Jüttler B (2007) Evolution-based least-squares fitting using Pythagorean hodograph spline curves. Comput Aided Geom Des 24: 310–322CrossRefGoogle Scholar
  3. 3.
    Al-Subaihi I, Watson GA (2004) The use of the l 1 and l norms in fitting parametric curves and surfaces to data. Appl Numer Anal Comput Math 1: 363–376zbMATHCrossRefMathSciNetGoogle Scholar
  4. 4.
    Alhanaty M, Bercovier M (2001) Curve and surface fitting and design by optimal control methods. Comput Aided Des 33: 167–182CrossRefGoogle Scholar
  5. 5.
    Atieg A, Watson GA (2003) A class of methods for fitting a curve or surface to data by minimizing the sum of squares of orthogonal distances. J Comput Appl Math 158: 277–296zbMATHCrossRefMathSciNetGoogle Scholar
  6. 6.
    Blake A, Isard M (1998) Active contours. Springer, New YorkGoogle Scholar
  7. 7.
    Boggs PT, Byrd RH, Schnabel RB (1987) A stable and efficient algorithm for nonlinear orthogonal distance regression. J Sci Stat Comput 8(6): 1052–1078zbMATHCrossRefMathSciNetGoogle Scholar
  8. 8.
    Bube KP, Langan RT (1997) Hybrid 1/ 2 minimization with application to tomography. Geophysics 62: 1183–1195CrossRefGoogle Scholar
  9. 9.
    Holland P, Welsch R (1977) Robust regression using iteratively re-weighted least-squares. Commun Stat Theory Methods 9(A6): 813–827Google Scholar
  10. 10.
    Hoschek J, Lasser D (1993) Fundamentals of computer aided geometric design. A K Peters, WellesleyzbMATHGoogle Scholar
  11. 11.
    Huber PJ (1981) Robust statistics. Wiley, New YorkzbMATHCrossRefGoogle Scholar
  12. 12.
    Jüttler B (1998) Computational methods for parametric discrete 1 and curve fitting. Int J Shape Model 4: 21–34zbMATHCrossRefGoogle Scholar
  13. 13.
    Kelley CT (1999) Iterative methods for optimization. Society for Industrial and Applied Mathematics, PhiladelphiaGoogle Scholar
  14. 14.
    Liu Y, Wang W (2008) A revisit to least squares orthogonal distance fitting of parametric curves and surfaces. In: Chen F, Jüttler B (eds) Advances in geometric modeling and processing, GMP 2008, Springer, pp 384–397Google Scholar
  15. 15.
    Mahadevan V, Narasimha-Iyer H, Roysam B, Tanenbaum HL (2004) Robust model-based vasculature detection in noisy biomedical images. IEEE Trans Inf Technol Biomed 8(3): 360–376CrossRefGoogle Scholar
  16. 16.
    McCullagh P, Nelder J (1998) Generalized linear models. Chapman & Hall, LondonGoogle Scholar
  17. 17.
    Osborne MR (1985) Finite algorithms in optimization and data analysis. Wiley, New YorkzbMATHGoogle Scholar
  18. 18.
    Pottmann H, Leopoldseder S, Hofer M, Steiner T, Wang W (2005) Industrial geometry: recent advances and applications in CAD. Comput Aided Des 37: 751–766CrossRefGoogle Scholar
  19. 19.
    Rogers D, Fog N (1989) Constrained B-spline curve and surface fitting. Comput Aided Des 21: 641–648zbMATHCrossRefGoogle Scholar
  20. 20.
    Sarkar B, Menq CH (1991) Parameter optimization in approximating curves and surfaces to measurement data. Comput Aided Geom Des 8: 267–280zbMATHCrossRefMathSciNetGoogle Scholar
  21. 21.
    Speer T, Kuppe M, Hoschek J (1998) Global reparametrization for curve approximation. Comput Aided Geom Des 15: 869–877zbMATHCrossRefMathSciNetGoogle Scholar
  22. 22.
    Wang W, Pottmann H, Liu Y (2006) Fitting B-spline curves to point clouds by curvature-based squared distance minimization. ACM Trans Graph 25(2): 214–238CrossRefGoogle Scholar
  23. 23.
    Watson GA (2000) Approximation in normed linear spaces. J Comput Appl Math 121: 1–36zbMATHCrossRefMathSciNetGoogle Scholar
  24. 24.
    Watson GA (2001) On the Gauss–Newton method for 1 orthogonal distance regression. IMA J Num Anal 22: 345–357CrossRefGoogle Scholar
  25. 25.
    Wedderburn RWM (1974) Quasi-likelihood functions, generalized linear models, and the Gauss– Newton method. Biometrika 61(3): 439–447zbMATHMathSciNetGoogle Scholar
  26. 26.
    Yamamoto T (2000) Historical developments in convergence analysis for Newton’s and Newton-like methods. J Comput Appl Math 124(1/2): 1–23zbMATHCrossRefMathSciNetGoogle Scholar

Copyright information

© Springer-Verlag 2009

Authors and Affiliations

  1. 1.Institute of Applied GeometryJohannes Kepler UniversityLinzAustria

Personalised recommendations