Optimal quantization of noisy sources

Problems in optimal multidimensional quantization of noise-corrupted sources are addressed. The optimal quantization rule, properties of the optimal quantizer, and its relations with the optimal estimator for the weighted squared error distortion measure are derived. An equivalent quantization technique is developed and the optimum codewords are calculated for the Gaussian signal in the additive independent Gaussian noise problem. Simulation results are presented.