Extrapolation of Incomplete Image Data with Discrete Orthogonal Transforms

Abstract

In image processing and transmission, interpolation and extrapolation are of great importance whenever missing pixels have to be filled in, and many methods have been proposed to solve this problem. In this paper we present a method for extrapolating the missing data with an existing set of basis functions of a selected orthogonal transform. The best extrapolation is found according to linear approximation theory as a weighted sum of basis functions, where coefficients of the sum are solutions of the derived matrix equation.