Nonlinear Image Interpolation Using Adaptive Conic Blending Spline

A new adaptive rational image interpolation algorithm is proposed. For a natural image, according to the difference information of the local pixel gray value in the vertical and horizonal direction, a conic blending spline with large weight parameters is employed to reconstruct sharp edges and image details within the interpolated image, and one with small weight is used to reduced the blocking artifacts. The proposed algorithm exhibits significant improvement in image quality when compared with the conventional polynomial spline type algorithms, especially at high resolution, such as four times or more.