Spatial error concealment method based on POCS with a correlation-based initial block

Abstract

A spatial error concealment method is proposed based on projections onto convex sets (POCS) with a correlation-based initial block. The conventional POCS results heavily depend on an initial block. To cope with this problem, a part of correctly received coefficients is utilised to find the most correlated neighbouring block, and a proper initial block is estimated from the found neighbouring block. The experiment shows that the proposed method has a much better performance.