Multiple Description Image and Video Coding for Noisy Channels

Abstract : This PhD thesis provides a new ways of looking multiple description methods (MDC).
We propose a robust joint source channel MDC method that is adapted to the transmission of images and videos on non stationary channels. The main advantage of this method is that it is well adapted for applications implying transmissions using unreliable and variable channels. Moreover, for its compression and synchronization capabilities it is appropriate for real time transmissions.
We show that the proposed method present the best results, in terms of signal to noise ratio and of visual quality, when compared with other multiple description methods existing in the state of the art. Moreover, it is well adapted for applications where the standard methods of error control cannot easily adapt to the characteristics of the channel.
The method is validated on various channel models (BSC, AWGN, Internet, UMTS, satellite) when considering image and video transmissions.
The proposed MDBA is well adapted for application demanding a feedback-based encoder, as peer-to-peer video conferencing, wireless video, etc, but it is insufficient in a real video streaming system where the server may be serving thousands of clients simultaneously. It is that the reason why we develop a system based in the MDBA but dedicated to video streaming. The last chapter of this dissertation presents an extension of the MDBA method to make streaming video robust to channel errors. The proposed method is rate adaptable, as the actual commercial methods, and it is also quality adaptable that is presented in the literature as an important characteristic of video streaming systems.
Document type :
Theses
Liste complète des métadonnées

https://tel.archives-ouvertes.fr/tel-00210328
Contributor : Estelle Nivault <>
Submitted on : Monday, January 21, 2008 - 2:12:26 PM
Last modification on : Wednesday, January 31, 2018 - 10:24:04 AM
Document(s) archivé(s) le : Wednesday, April 14, 2010 - 10:10:20 PM

Identifiers

  • HAL Id : tel-00210328, version 1

Collections

Citation

Manuela Pereira. Multiple Description Image and Video Coding for Noisy Channels. Automatic. Université Nice Sophia Antipolis, 2004. English. ⟨tel-00210328⟩

Share

Metrics

Record views

447

Files downloads

631