A New Method by Rational Interpolation for Image Zooming
Min Sheng · Computer Knowledge and Technology · 2010
The paper presents a new method for image zooming by reconstructing the kernal functions of image interpolation via multivariate blending osculatory rational interpolants since that the rational functions can approximate the large deflection surfaces. This algorithm can effectively preserve the characters of image with the location and direction on the edge areas. Furthermore, the paper regards the pixel of color image as a pure quaternion and presents quaternion continued fraction interpolants by the theories of Clifford algebra. The method will possess efficient computation and verify that the more rich details on the edges of the color image have greater improvements than the cases of traditional methods using quaternion interpolation method.