DocumentCode :
970353
Title :
On the support of fixed-rate minimum mean-squared error scalar quantizers for a Laplacian source
Author :
Na, Sangsin
Author_Institution :
Sch. of Electr. & Comput. Eng., Ajou Univ., Suwon, South Korea
Volume :
50
Issue :
5
fYear :
2004
fDate :
5/1/2004 12:00:00 AM
Firstpage :
937
Lastpage :
944
Abstract :
This correspondence shows that the support growth of a fixed-rate optimum (minimum mean-squared error) scalar quantizer for a Laplacian density is logarithmic with the number of quantization points. Specifically, it is shown that, for a unit-variance Laplacian density, the ratio of the support-determining threshold of an optimum quantizer to 3/√2lnN/2 converges to 1, as the number N of quantization points grows. Also derived is a limiting upper bound that says that the support-determining threshold cannot exceed the logarithmic growth by more than a small constant, e.g., 0.0669. These results confirm the logarithmic growth of the optimum support that has previously been derived heuristically.
Keywords :
Laplace equations; least mean squares methods; quantisation (signal); Laplacian source density; asymptotic quantization; fixed-rate minimum mean-squared error scalar quantizers; log-linearity; support region; support-determining threshold; Laplace equations; Probability density function; Quantization; Random variables; Source coding; Upper bound;
fLanguage :
English
Journal_Title :
Information Theory, IEEE Transactions on
Publisher :
ieee
ISSN :
0018-9448
Type :
jour
DOI :
10.1109/TIT.2004.826686
Filename :
1291745
Link To Document :
بازگشت