Computer Vision

Living Edition

Optimal Estimation

  • Kenichi KanataniEmail author
Living reference work entry


Related Concepts


Optimal estimation in the computer vision context refers to estimating the parameters that describe the underlying problem from noisy observation. The estimation is done according to a given criterion of optimality, for which maximum likelihood is widely accepted. If Gaussian noise is assumed, it reduces to minimizing the Mahalanobis distance. If furthermore the Gaussian noise has a homogeneous and isotropic distribution, the procedure reduces to minimizing what is called the reprojection error.


One of the central tasks of computer vision is the extraction of 2D/3D geometric information from noisy image data. Here, the term image data refers to values extracted from images by image processing operations such as edge filters and interest point detectors. Image data are said to be noisyin the sense that...

This is a preview of subscription content, log in to check access.


  1. 1.
    Chernov N, Lesort C (2004) Statistical efficiency of curve fitting algorithms. Comput Stat Data Anal 47(4):713–728MathSciNetCrossRefGoogle Scholar
  2. 2.
    Kanatani K (2008) Statistical optimization for geometric fitting: theoretical accuracy analysis and high order error analysis. Int J Comput Vis 80(2):167–188MathSciNetCrossRefGoogle Scholar
  3. 3.
    Neyman J, Scott EL (1948) Consistent estimates based on partially consistent observations. Econometrica 16(1):1–32MathSciNetCrossRefGoogle Scholar
  4. 4.
    Kanatani K, Sugaya Y, Kanazawa Y (2016a) Ellipse fitting for computer vision: implementation and applications. Morgan Claypool, San RafaelGoogle Scholar
  5. 5.
    Kanatani K, Sugaya Y, Kanazawa Y (2016b) Guide to 3D vision computation: geometric analysis and implementation. Springer, ChamCrossRefGoogle Scholar
  6. 6.
    Sampson PD (1982) Fitting conic sections to “very scattered” data: an iterative refinement of the Bookstein algorithm. Comput Graphics Image Process 18(1):97–108CrossRefGoogle Scholar
  7. 7.
    Chojnacki W, Brooks MJ, van den Hengel A, Gawley D (2000) On the fitting of surfaces to data with covariances. IEEE Trans Patt Anal Mach Intell 22(11):1294–1303CrossRefGoogle Scholar
  8. 8.
    Leedan Y, Meer P (2000) Heteroscedastic regression in computer vision: problems with bilinear constraint. Int J Comput Vis 37(2):127–150CrossRefGoogle Scholar
  9. 9.
    Kanatani K, Sugaya Y (2010) Unified computation of strict maximum likelihood for geometric fitting. J Math Imaging Vis 38(1):1–13MathSciNetCrossRefGoogle Scholar
  10. 10.
    Taubin G (1991) Estimation of planar curves, surfaces, and non-planar space curves defined by implicit equations with applications to edge and range image segmentation. IEEE Trans Patt Anal Mach Intell 13(11):1115–1138CrossRefGoogle Scholar
  11. 11.
    Kanatani K, Rangarajan P, Sugaya Y, Niitsuma H (2011) HyperLS for parameter estimation in geometric fitting. IPSJ Trans Comput Vis Appl 3:80–94CrossRefGoogle Scholar
  12. 12.
    Kanatani K (1993) Renormalization for unbiased estimation. In: Proceedings of 4th international conference on computer vision, Berlin, pp 599–606Google Scholar
  13. 13.
    Kanatani K, Al-Sharadqah A, Chernov N, Sugaya Y (2014) Hyper-renormalization: non-minimization approach for geometric estimation. IPSJ Trans Comput Vis Appl 6:143–149CrossRefGoogle Scholar
  14. 14.
    Kanatani K, Sugaya Y (2013) Hyperaccurate correction of maximum likelihood for geometric estimation. IPSJ Trans Comp Vis Appl 5:19–29CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2020

Authors and Affiliations

  1. 1.Professor EmeritusOkayama UniversityOkayamaJapan

Section editors and affiliations

  • Koichiro Deguchi
    • 1
  1. 1.Tohoku UniversitySendaiJapan