Title :
On the structure of entropy-constrained scalar quantizers
Author :
György, András ; Linder, Tamás
Author_Institution :
Dept. of Comput. Sci. & Inf. Theory, Tech. Univ. Budapest, Hungary
Abstract :
New results concerning the structure and existence of optimal entropy-constrained scalar quantizers (ECSQs) are presented. One main result shows that for continuous sources and a wide class of distortion measures, any finite-level ECSQ can be replaced by an ECSQ with interval cells which has the same entropy and equal or less distortion. Furthermore, the existence of an optimal ECSQ for an arbitrary entropy constraint is shown under rather general conditions
Keywords :
entropy codes; quantisation (signal); rate distortion theory; source coding; ECSQ; arbitrary entropy constraint; continuous sources; distortion measures; interval cells; optimal entropy-constrained scalar quantizers; rate-distortion function; Computer science; Distortion measurement; Entropy; Information theory; Nearest neighbor searches; Pathology; Q measurement; Quantization; Random variables; Tin;
Conference_Titel :
Information Theory, 2001. Proceedings. 2001 IEEE International Symposium on
Conference_Location :
Washington, DC
Print_ISBN :
0-7803-7123-2
DOI :
10.1109/ISIT.2001.935892
