Color Image Magnification with Reserved Sharp Contour Based on Rational-Linear Interpolation
Jieqing Tan · Acta Simulata Systematica Sinica · 2004
A novel adaptive interpolation magnification algorithm for color image is presented, which is based on rational-linear vector valued interpolation. In this case, the RGB values of an image pixel are regarded as the components of a vector. First of all, we divide a given image into overlapping N譔 blocks according to the proposed scheme, analyze the local characteristics of the blocks, construct Newton-Thiele抯 rational interpolation surfaces or Newton-Newton抯 linear interpolation surfaces, and then, resample the interpolation surfaces with different sampling rates to generate the required resized image. Experimental results show that the algorithm is superior to that of classical linear interpolation schemes.