Image interpolation using adaptive P-spline

This paper introduces a novel scheme for image interpolation. The basic contribution of the proposed interpolation scheme lies in two aspects. Firstly, the missing pixels in the image to be interpolated are classified into smooth and non-smooth based on the local characteristic of the image. Secondly regularized version of B-spline i.e. penalized spline (P-spline) interpolation scheme is used for interpolating the missing pixel of the image. A weighted penalty is incorporated in P-spline interpolation which adaptively penalizes the classified image regions and gives proper emphasis to each individual missing pixel. The proposed interpolation scheme is tested for various natural and synthetic images. Experimental results reveal the supremacy of the suggested interpolation scheme compared with the other prevailing state of art interpolation methods.