Globally Optimal Vector Quantizer Design by
K. Zeger, Jacques Vaisey, A. Gersho · 1992
This paper presents a unified formulation and study of vector quantizer design methods that couple stochastic re- laxation (SR) techniques with the generalized Lloyd algorithm. Two new SR techniques are investigated and compared: simu- lated annealing (SA), and a reduced-complexity approach that modifies the traditional acceptance criterion for simulated an- nealing to an unconditional acceptance of perturbations. It is shown that four existing techniques all fit into a general meth- odology for vector quantizer design aimed at finding a globally optimal solution. Comparisons of each algorithms' perfor- mance when quantizing Gauss-Markov processes, speech, and image sources are given. The SA method is guaranteed to per- form in a globally optimal manner, and the SR technique gives empirical results equivalent to those of SA. Both techniques re- sult in significantly better performance than that obtained with the generalized Lloyd algorithm.