An adaptive irregular sampling algorithm and its application to image coding

A novel irregular and nonuniform sampling scheme for image data is presented in this paper. Its main characteristics are the particular distribution of the samples along image edges and in textured areas, which is obtained following a multiresolution approach, and the low computational complexity. As an example of application, an image coding system is proposed which, thanks to both the peculiar sample distribution and a suitably designed interpolation scheme, yields good subjective quality images at low bit rates, avoiding the annoying artefacts which are typical of block-based coding techniques as JPEG. (C) 2001 Elsevier Science B.V. All rights reserved.