Rioul, Olivier (1993) Regular wavelets: application to still image compression. PhD thesis Signal et Images, Signal, ENST TELECOM Paris 93 E 004.
Full text available as:
|
|
Alternative Locations: http://perso.enst.fr/~rioul/pdf/19930315rioulthese.pdf
Abstract
This thesis, started at the end of 1989, is devoted to the study of the effect of new mathematical properties brought by wavelet theory on subband coding, for still image compression problems.
Using a theoretical analysis of discrete-time multiresolution decomposition, it is shown that, within our framework, the practical novelty of wavelets reduces to the {\em regularity\/}property of octave-band filter banks used in subband coding schemes.
Then, several tools are derived, which make a systematic study of regularity possible, for applications such as image compression.
A mathematical study of the regularity property is investigated,in which optimal regularity estimation algorithms are derived.
Several filter bank design algorithms are then proposed.
They allow one to balance regularity against other filter properties (orthogonality, linear phase response, frequency selectivity, etc.) which are generally believed to be relevant for image coding.
Fast algorithms are also derived, for which substancial reduction of computational load required by octave-band filter banks is possible.
Finally, an experimental study of filter properties is carried out, for a simple still image compression scheme, in which the wavelet transform is separable.
In this restricted framework, the potential interest of regularity is emphasized, in competition with other criteria such as frequency selectivity and linear phase.
| Item Type: | PhD Thesis (PhD) |
|---|---|
| Additional Information: | This thesis was carried out while the author was with the Centre National d'Etudes des Telecommunications (CNET), Issy-les-Moulineaux, France. |
| Thesis Supervisor: | Duhamel, Pierre |
| Date: | March 1993 |
| Board of examiners: | Labit, Claude and Vetterli, Martin and Meyer, Yves and Barlaud, Michel and Grenier, Yves |
| Ecole Doctorale: | ED 130 INFORMATIQUE, TELECOMMUNICATIONS ET ELECTRONIQUE (EDITE) |
| Discipline: | Signal et Images |
| Collection (Fonds): | ENST |
| Institution: | ENST |
| Department: | Signal |
| Subjects: | 2. Information and Communication Sciences and Technologies |
| Uncontrolled Keywords: | Wavelets regularity compression image, Ondelettes régularité compression images |
References
O. Rioul, Ondelettes Réulières: Application à la Compression d'Images Fixes, PhD Thesis, ENST, Télécom Paris 93 E 004, Spécialité: Signal et Images, 259 pages, Mar. 1993.
Table of content
1. Introduction
2. Ondelettes discrètes
3. Ondelettes continues
4. Régularité
5. Calcul de bancs de filtres réguliers
6. Algorithmes rapides
7. Schéma de compression d'images fixes
8. Résultats
9. Conclusion générale et perspectives
A. Articles
B. Bibliographie
C. Boîte à outils MATLAB
| ID Code: | 1307 |
|---|---|
| Deposited By: | Olivier Rioul |
| Deposited On: | 01 July 2005 |
Repository Staff Only: edit this item