Structured Vector Quantization

本项目涉及的理论和设计， 结构化矢量量化器的压缩和存储 模拟源，例如语音、图像、视频和音乐。 向量 量化是一类技术， 与最近的压缩技术相比，有希望获得显著的收益。 矢量量化器的困难在于高度的 实现中涉及的复杂性。本项目的重点 是一个理论的发展，减少复杂性向量 量化器。 渐近量子化理论是进一步 开发并用于分析性能的中度到高度 速率结构量化器。 新的分析方法 正在研究低速率量化器的性能。 的 性能与结构的基本限制 探讨了

This project is concerned with the theory and design of structured vector quantizers for the compression and storage of analog sources such as speech, images, video, and music. Vector quantization is a class of techniques that holds significant promise for significant gains over recent compression techniques. The difficulty with vector quantizers is the high degree of complexity involved in implementations. The focus of this project is the development of a theory of reduced complexity vector quantizers. Asymptotic quantization theory is to be further developed and used to analyze the performance of moderate to high rate structured quantizers. New methods for analyzing the performance of low rate quantizers are being studied. The fundamental limits of performance versus structure are being explored.