Adaptive osculatory rational interpolation for image processing

Image interpolation is a common problem in image applications. Although many interpolation algorithms have been proposed in the literature, these methods suffer from the effects of imperfect reconstruction to some degree, most often, these effects manifest themselves as jagged contours or blurred edges in the image. This paper presents a method for preserving the contours or edges based on adaptive osculatory rational interpolation kernel function, which is built up by approximating the ideal interpolating kernel function by continued fractions. It is a more accurate approximation for the ideal interpolation in space domain or frequency domain than by other linear polynomial interpolation kernel functions. Simulation results are also presented to demonstrate the superior performance of image magnification.