Ellipse Fitting with Hyperaccuracy
For fitting an ellipse to a point sequence, ML (maximum likelihood) has been regarded as having the highest accuracy. In this paper, we demonstrate the existence of a “hyperaccurate” method which outperforms ML. This is made possible by error analysis of ML followed by subtraction of high-order bias terms. Since ML nearly achieves the theoretical accuracy bound (the KCR lower bound), the resulting improvement is very small. Nevertheless, our analysis has theoretical significance, illuminating the relationship between ML and the KCR lower bound.
KeywordsThick Solid Line Thin Solid Line Order Error Point Sequence True Shape
- 11.Kanatani, K.: Statistical Optimization for Geometric Computation: Theory and Practice. Elsevier Science, Amsterdam (1996); reprinted by Dover, New York (2005)Google Scholar
- 13.Kanatani, K.: Further improving geometric fitting. In: Proc. 5th Int. Conf. 3-D Digital Imaging and Modeling (3DIM 2005), Ottawa, Ontario, Canada, June 2006, pp. 2–13 (2006)Google Scholar
- 14.Kanazawa, Y., Kanatani, K.: Optimal conic fitting and reliability evaluation. IEICE Trans. Inf. & Sys. E79-D(9), 1323–1328 (1996)Google Scholar
- 19.Mühlich, M., Mester, R.: Unbiased errors-in-variables estimation using generalized eigensystem analysis. In: Proc. 2nd Workshop Statistical Methods in Video Processing, Prague, Czech, May 2004, pp. 38–49 (2004)Google Scholar