par Cardinal, Jean 
Référence Journal of VLSI Signal Processing Systems for Signal, Image, and Video Technology, 33, 3, page (287-294)
Publication Publié, 2003-03
Référence Journal of VLSI Signal Processing Systems for Signal, Image, and Video Technology, 33, 3, page (287-294)
Publication Publié, 2003-03
Article révisé par les pairs
| Résumé : | Multiple description coding aims at transmitting two mutually refinable descriptions of a signal on two different communication channels. It is successfully applied to transmission of sound and images on diversity systems or packet-switched networks. We present novel algorithms for the design of multiply descriptive tree-structured vector quantizers with linear encoding complexity that improve on our previously published methods. These algorithms allow for both the rates and distortions constraints in a natural way and are shown to yield quantizers that are competitive with full-search codes in terms of rate-distortion performance. We also present a generalization of the scheme to more than two channels. |
