分布式信源矢量量化性能分析

Performance Analysis for Vector Quantization of Distributed Source

研究分布式信源量化维数与量化性能之间的定量关系.通过量化器输出点密度函数分别描述了量化码率和失真,推导出量化码率、失真、维数之间的关系,并分析了量化渐近性能.得到量化维数同量化性能之间的关系函数.仿真结果表明,随着维数增加,量化率失真性能渐进逼近Wyner-Ziv限.在较低维数情况下,分布式信源矢量量化也能较好地逼近Wyner-Ziv限,即可以用低复杂度的低维量化器达到性能的需求.

To study the quantitative relation of dimension and complexity. The rate and distortion of distributed source vector quantization are characterized and analyzed through the point density function of a quantizer. The function of rate, distortion, dimension is then derived, and the asymptotic performance of vector quantization is analyzed. The function of dimension and performance is then obtained. Simulation results showed that rate-distortion performance can approach Wyner-Ziv bound gradually with increasing dimension. The performance of distributed source vector quantization can approach Wyner-Ziv bound in the lower dimension case. So performance requirements can be met by low-dimension quantizer of less complexity.