Structured Vector Quantization

结构化矢量量化

• 批准号: 9105647
• 负责人: David Neuhoff
• 金额: $ 17.34万
• 依托单位: Regents of the University of Michigan - Ann Arbor
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1991
• 资助国家: 美国
• 起止时间: 1991-10-01 至 1994-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9105647&HistoricalAwards=false
• 关键词: Structured; Vector; Quantization

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.

该项目关注结构化矢量量化器的理论和设计，用于压缩和存储模拟源，如语音，图像，视频和音乐。向量量化是一种技术，与最近的压缩技术相比，它有很大的发展前景。矢量量化器的难点在于其实现的高度复杂性。这个项目的重点是开发一个降低复杂度的矢量量化器理论。渐近量化理论有待进一步发展，并用于分析中高速率结构化量化器的性能。人们正在研究分析低速率量化器性能的新方法。人们正在探索性能与结构之间的基本界限。