Skip to Main content Skip to Navigation

Une approche vectorielle pour exploiter le contenu de l'image en compression et tatouage

Abstract : The first part of this document deals with still image compression scheme using discrete wavelets transform (DWT) and lattice vector quantization (LVQ). Two vector quantizer have been proposed: a dead zone lattice vector quantizer and a reduced-lattice vector quantizer. In addition, two algorithms have been designed in order to reduce the complexity of the proposed compression schemes : an optimal (in the distortion sense) dead zone search and a rate allocation based on an exponential approximation of the rate-distortion functions. In the second part, two joint compression and watermarking schemes, based on DWT and LVQ, are presented. The first method takes into account the image content when embedding the mark and the second, called modulation lattice vector quantization (MLVQ), is based on quantization index modulation principles. In particular, we have shown that MLVQ must be added to a vector dead zone and a specific indexing method to maintain good compression performances. Finally, a quantization parameter estimation method has been proposed in order to allow variable rate coding.
Document type :
Complete list of metadata
Contributor : Thèses UL Connect in order to contact the contributor
Submitted on : Thursday, March 29, 2018 - 10:43:00 AM
Last modification on : Saturday, October 16, 2021 - 11:14:14 AM

Links full text


  • HAL Id : tel-01746704, version 1



Ludovic Guillemot. Une approche vectorielle pour exploiter le contenu de l'image en compression et tatouage. Autre. Université Henri Poincaré - Nancy 1, 2004. Français. ⟨NNT : 2004NAN10229⟩. ⟨tel-01746704⟩



Record views