Algorithmic Theory of Universal Source Coding with a Fidelity Criterion and Related Topics

文章摘要：美国南加州大学NCR-9508282，《基于保真度准则的通用信源编码算法理论及相关课题》，皮张：该项目解决的问题是：能否为有损信源编码设计一种在线的通用顺序信源编码算法？在线通用顺序有损算法的含义是这样的算法是顺序的，一过率失真限值，延迟短，自适应性好，编码复杂度相对较低，适合在线实现。将有必要研究通用有损信源编码算法的基本特性，并且这将需要开发用于算法分析的系统方法，例如分析算法的冗余，对算法和其他算法的健壮性和编码复杂性的分析topics.*****************************************************************************Aubrey M.BushProgram主任，代理副主任，网络和通信研究与基础设施司

ABSTRACT: NCR-9508282, University of Southern California, "Algorithmic Theory of Universal Source Coding with a Fidelity Criterion and Related Topics", PI-Zhang: This project addresses the question "Can one design an on-line universal sequential source coding algorithm for lossy source coding?" What is meant by an on-line universal sequential lossy algorithm is that such an algorithm is sequential, achieves in one pass the rate-distortion limit, and has a short delay, good adaptivity, and relatively low coding complexity, so that it is suitable for on-line implementation. It will be necessary to study the basic properties of a universal lossy source coding algorithm, and this will require the development of systematic methods for algorithm analysis such as the analysis of the redundancy of an algorithm, the analysis of the robustness and coding complexity of an algorithm and other topics.*****************************************************************************Aubrey M. BushProgram Director, Acting Deputy Divison DirectorDivision of Networking and Communications Research and InfrastructureNational Science Foundation