Sign96

J. Signes. Reducing the codebook size in fractal image compression by geometrical analysis. In Proc. SPIE VCIP 1996, Volume 2727, 1996.

Abstract

In most IFS based image coding schemes, the domain blocks codebook searched is independent from the input image, identical for all range blocks of the same size, and non optimized. In order to get a good match between range and domain blocks, a huge codebook is used, which is both inefficient in terms of computational complexity abd output bit rate. We propose to design an optimal reduced codebook for the whole image thanks to a geometrical analysis of the basic scheme. We then compare this method with a "local" codebook method which leads us to some conclusions about the differences between IFS and vectorial quantization (VQ) based coding schemes.

BibTex Reference

@InProceedings{Sign96,
   Author = {Signes, J.},
   Title = {Reducing the codebook size in fractal image compression by geometrical analysis},
   BookTitle = {Proc. SPIE VCIP 1996},
   Volume = {2727},
   Year = {1996}
}