Analysis of stationary subdivision schemes for curve design based on radial basis function interpolation (Q2266961)

The authors provide a large family of interpolatory stationary subdivision schemes based on radial basis functions which are positive definite or conditionally positive definite. Well known examples are multiquadrics \(\varphi_\lambda(x) = (x^2+\lambda^2)^{\beta/2}\), Gaussians \(\varphi_\lambda(x)= \exp(-|x|^2/\lambda^2)\) and inverse multiquadrics \(\varphi_\lambda(x) =(x^2+\lambda^2)^{-1/2}\). The radial basis functions have a tension parameter \(\lambda >0\) and it is proven that for a sufficiently large \(\lambda \geqslant \lambda _0\), the proposed \(2L\)-point \((L\in \mathbb N)\) scheme has the same smoothness as the well-known \(2L\)-point Deslauriers-Dubuc scheme [\textit{G. Deslauriers} and \textit{S. Dubuc}, Constructive Approximation 5, No. 1, 49--68 (1989; Zbl 0659.65004)], which is based on \(2L-1\) degree polynomial interpolation. Some numerical examples using the three examples of radial basis functions are presented to illustrate the performance of the new schemes, adapting subdivision rules on bounded intervals in a way of keeping the same smoothness and accuracy of the pre-existing schemes on \(\mathbb R\). With proper tension parameters, the new scheme can alleviate undesirable artifacts near boundaries, which usually appear to interpolatory schemes with irregularly distributed control points.