Linear scalar quantization of wavelet image decomposition using joint optimization
Riadh Abdelfattah, Olivier Rioul, Pierre Duhamel · 2005
This paper proposes a new linear scalar quantization algorithm. A modified Lloyd's optimum algorithm, based on a linear quantization, is proposed to decompose a memory-less source into the bits. This is the case of the considered source which is the output of the wavelet transformation. The developed algorithm gives acceptable results for generalized Gaussian sources. Our approach is compared with the Lloyd max one's and a second one where the quantization intervals are equiprobable.