On the Coupling of Decimation Operator with Subdivision Schemes for Multi-scale Analysis
Subdivision schemes [5, 11] are powerful tools for the fast generation of refined sequences ultimately representing curves or surfaces. Coupled with decimation operators, they generate multi-scale transforms largely used in signal/image processing [1, 3] that generalize the multi-resolution analysis/wavelet framework . The flexibility of subdivision schemes (a subdivision scheme can be non-stationary, non-homogeneous, position-dependent, interpolating, approximating, non-linear...) (e.g. ) is balanced, as a counterpart, by the fact that the construction of suitable consistent decimation operators is not direct and easy.
In this paper, we first propose a generic approach for the construction of decimation operators consistent with a given linear subdivision. A study of the so-called prediction error within the multi-scale framework is then performed and a condition on the subdivision mask to ensure a fast decay of this error is established. Finally, the cases of homogeneous Lagrange interpolatory subdivision, spline subdivision, subdivision related to Daubechies scaling functions (and wavelets) and some recently developed non stationary non interpolating schemes are revisited.
KeywordsMulti-scale analysis Decimation Subdivision
- 5.Cavaretta, A., Dahmen, W., Micchelli, C.: Stationary subdivision. In: Memoirs of the American Mathematics Society, vol. 93, No. 453, Providence, Rhode Island (1991)Google Scholar
- 9.De Rham, G.: Un peu de mathématiques à propos d’une courbe plane. Elem. Math. 2, 73–76 (1947)Google Scholar
- 10.Deslauries, G., Dubuc, S.: Interpolation dyadique. In: Fractals, dimensions non entières at applications, pp. 44–55 (1987)Google Scholar
- 11.Dyn, N.: Subdivision schemes in computer-aided geometric design. In: Light, W. (ed.) Advances in Numerical Analysis II, Wavelets, Subdivision Algorithms and Radial Basis Functions, pp. 36–104. Clarendon Press, Oxford (1992)Google Scholar
- 13.Kui, Z.: Approximation multiechelle non lineaire et applications en analyse de risques. Ph.D. thesis, Ecole Centrale Marseille, Marseille, France (2018)Google Scholar