Contour matching using wavelet transform and multigrid methods
In this paper, wavelet transform and multigrid method are combined to make the method more practical. It is known that Gaussian filtering causesshrinkage of data. To overcome this disadvantage, Gaussian filtering is replaced with wavelet transform. This method introduces no curve shrinkage. Then, the linearized form of objective equation is proposed. This makes contour matching easier to implement. Finally, the multigrid method is used to speed up the convergence.
KeywordsContour matching curvature wavelet transform multigrid method
- Mokhtarian F, Mackworth A. Scale-based description and recognition of planar curves and two-dimension shapes.IEEE Trans. Pattern Anal. Machine Intell., 1986, PAMI-8(1): 34–43.Google Scholar
- Witkins A. A ‘Scale-space filtering’. InProc. 7th Int’l Joit Conf. Artif. Intell., Kalsruhe, Germany, 1983.Google Scholar
- Yang X, DeMuynck P, Nyssen Eet al. A multiscale finite difference method for contour matching. InThe Third Int’l Conf. Automation, Robotics and Computer Vision (ICARCV’94), Singapore, 1994, pp. 2010–2014.Google Scholar
- Daubechies I. Ten lectures on wavelets.CBMS-NSF Series in Appl. Math., SIAM, Philadelphia, 1991.Google Scholar
- Hackbusch W. Multigrid Methods and Applications. Springer-Verlag, Berlin, 1985.Google Scholar