Recursive minimum energy algorithm for image interpolation

Abstract

An interpolation method based on the criterion to maximize the smoothness is proposed. In contrast to the conventional approach, where the interpolated signal is made smooth by low-pass filtering, in the proposed method, the smoothness is obtained by minimizing a global energy function. With the minimum energy formulation for interpolation, a recursive method is used to derive the solution and named as the recursive minimum energy algorithm. Because the smoothness is improved by increasing the calculation iteration but not by increasing the tap length, the controllability of this system is increased. Experiments show that the smoothness and the frequency performance obtained by the proposed algorithm are better than those of the conventional method. Also, the corresponding architecture design for the image interpolation is presented and shown to be suit for VLSI implemented.