Digital Image Interpolation Method Using Higher-Order Hermite Interpolating Polynomials with Compact Finite-Difference
Ryo Seta, Kan Okubo, Norio Tagawa · Institutional Repositories DataBase (IRDB) · 2009
Image interpolation can be performed by a convolution operation using the neighboring image values. To achieve accurate image interpolation, some of the conventional methods use basis function with large support, and therefore their implementation may have a large computational cost. Interpolation by the Hermite interpolating polynomials can be performed using image values and their derivatives. This makes it possible to realize the high-order interpolation with small support. In this study, we show that, by introducing the basis function of derivatives, interpolation methods by the higher-order Hermite interpolating polynomials can be expressed as a convolution form similarly to the conventional methods. Thus, the higher-order interpolation with small support is obtained as well. Moreover, high accuracy is achieved by using the compact FDs as the calculation method of derivatives. As a result, the efficiency of this method is confirmed, by comparing it with the conventional methods which have the same support.