Find a copy in the library
Finding libraries that hold this item...
|All Authors / Contributors:||
Salma Ben Jamaa-Joussemet; Michel Kieffer; Pierre Duhamel; Université de Paris-Sud. Faculté des Sciences d'Orsay (Essonne).
|Description:||1 vol. (XII-170 p.) : ill. ; 30 cm.|
|Responsibility:||Salma Ben Jamaa-Joussemet ; sous la direction de [Michel Kieffer et Pierre Duhamel].|
This thesis deals with joint source-channel coding and decoding (JSCC/JSCD) schemes involving arithmetic codes (AC). First, we develop a JSCD technique based on MAP estimation of CABAC encoded data. This estimator is considered to be exact as it is evaluated without approximations and with respect of the constraints imposed by an actual implementation of AC. The evaluation of the proposed MAP estimator is achieved using an improved sequential decoding technique, allowing to adjust the decoder complexity-efficiency trade-off. The purpose of the second part of this thesis is to provide analytical tools to predict and evaluate the effectiveness of the redundancy introduced by the JSCC schemes into AC. Integer binary AC is then modelled by a reduced-state automaton to obtain a bit-clock trellis. Distance spectra are then derived. The obtained distance properties provide an objective criteria that are then exploited to design efficient error-correcting arithmetic codes.