Skip to Main content Skip to Navigation
Theses

Compression d'images avec et sans perte par la méthode LAR (Locally Adaptive Resolution)

Abstract : The aim of this work consists in the definition of a unified method for lossy and lossless compression of still images or videos. The LAR method (Locally Adaptive Resolution) was first designed for lossy low bitrates coding. The self-extracting region representation provides efficient coding in terms of quality and bitrate. Three scalable lossless compression schemes are defined thanks to pyramidal structures: the LAR-APP exploits a refined prediction context, the Interleaved S+P interleaves two S-pyramids, the RWHT+P uses a new form of WHT 2D. Results are shown to be much better than state-of-the-art ones. Moreover, a joint LAR-Mojette source-channel coding is designed for secured transmission of medical images. Furthermore, an automatic implementation of LAR coders on multi-components heterogeneous parallel architectures is also realized. Finally, the extension of the LAR to the video is the subject of an exploratory work.
Complete list of metadatas

Cited literature [214 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00131758
Contributor : Marie Babel <>
Submitted on : Monday, February 19, 2007 - 12:09:03 PM
Last modification on : Wednesday, October 14, 2020 - 3:41:20 AM
Long-term archiving on: : Wednesday, April 7, 2010 - 12:12:42 AM

Identifiers

  • HAL Id : tel-00131758, version 1

Citation

Marie Babel. Compression d'images avec et sans perte par la méthode LAR (Locally Adaptive Resolution). Traitement du signal et de l'image [eess.SP]. INSA de Rennes, 2005. Français. ⟨tel-00131758⟩

Share

Metrics

Record views

1513

Files downloads

3797