Design of globally optimal multiple description scalar quantizers /

This thesis investigates the design of globally optimal multiple description scalar quantizers. The design of multiple description scalar quantizers is formulated as a combinatorial optimization problem due to the index assignment issue. The first approach uses the simulated annealing technique co...

Full description

Bibliographic Details
Main Author: Desilva, Buveneka Kanishka, 1964-
Format: Thesis eBook
Language:English
Published: [Place of publication not identified] : [publisher not identified] ; 1993.
Subjects:
Online Access:Link to OAKTrust copy
Description
Summary:This thesis investigates the design of globally optimal multiple description scalar quantizers. The design of multiple description scalar quantizers is formulated as a combinatorial optimization problem due to the index assignment issue. The first approach uses the simulated annealing technique coupled with the generalized Lloyd algorithm developed for multiple description scalar quantizers. The second approach utilizes deterministic annealing, where a sequence of smooth energy functions, parameterized by a temperature coefficient is generated. The design of the multiple description scalar quantizer with the simulated annealing algorithm resulted in good index assignments. For Gaussian and Laplacian sources, a significant improvement in the overall distortion of the quantizer was achieved. For the Uniform source, no improvement was observed. The deterministic annealing algorithm used for the design of the multiple description scalar quantizer, performed poorly and was worse than the conventional design method at high rates.
Item Description:"Major subject: Electrical Engineering".
Vita.
Physical Description:xi, 72 leaves : illustrations ; 28 cm.
Also available online.
Bibliography:Includes bibliographical references.