DocumentCode :
2466760
Title :
Optimal bit allocation for universal fixed-rate vector quantization
Author :
Kanlis, Angelos ; Narayan, Prakash
Author_Institution :
Found. for Res. & Technol., Inst. of Comput. Sci., Crete, Greece
fYear :
1998
fDate :
16-21 Aug 1998
Firstpage :
171
Abstract :
We investigate the performance of a mismatched fixed-rate lossy code which is designed for a source different from the “true” source. The notion of rate-distortion discrepancy between two pmf´s is shown to provide a lower bound for the expected excess rate due to mismatch for every fixed-rate code. This notion is used to study the problem of optimal rate allocation for the descriptions of source statistics and data in a two-stage universal vector quantisation scheme
Keywords :
rate distortion theory; source coding; vector quantisation; VQ; expected excess rate; lower bound; mismatched fixed-rate lossy code; optimal bit allocation; optimal rate allocation; rate-distortion discrepancy; source statistics; two-stage vector quantisation scheme; universal fixed-rate vector quantization; Bit rate; Block codes; Computer science; Distortion measurement; NASA; Performance loss; Rate-distortion; Statistics; Stochastic processes; Vector quantization;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Information Theory, 1998. Proceedings. 1998 IEEE International Symposium on
Conference_Location :
Cambridge, MA
Print_ISBN :
0-7803-5000-6
Type :
conf
DOI :
10.1109/ISIT.1998.708767
Filename :
708767
Link To Document :
بازگشت