DISCRETE ORTHOGONAL TRANSFORM FOR GAPPY IMAGE EXTRAPOLATION
Jaroslav Polec, T. Karlubikova · Kluwer Academic Publishers eBooks · 2006
An important issue in object-based image coding is the efficient description of image segments having an arbitrary shape, that is able to cope with missing pixels or blocks (for example present due to transmission errors). In this paper we present a method for approximating the missing data with an existing set of basis functions of selected orthogonal transform. Best extrapolation is found according linear approximation theory as weighted sum of basis functions, where coefficients of the sum are solutions of derived matrix equation.