【作者】 孙伟； 许君一； 齐东旭；

【Author】 Sun Wei Xu Junyi Qi Dongxu (School of Information Science and Technology, Zhongshan University, Guangzhou 510275) (Macao University of Science and Technology, Macao)

【机构】 中山大学信息科学与技术学院；

【摘要】 多结点样条(many-knot spline)函数是在B样条函数中引入更多附加结点构造出的一类具有有限支集的拉格朗日型样条函数。本文利用多结点样条拟合提出一种自适应的多层次淘汰冗余数据的方法,并通过不同的采样数据对算法进行了有效的论证。该方法充分利用了多结点样条函数拟合的基数型、显式计算(无需求解方程组)和局部性等优点,基于多尺度分析的思想,以自适应的方式分层次地去除数据的相关性,达到了用少数据量表示大量数据的目的。对于解决类似问题的思路来讲, 是一种新的尝试;可以直接用于数据压缩,或现有压缩算法的预处理(或后处理)。更多还原

【Abstract】 Many-knot spline function is a Lagrange type function with finite support, which is constructed by adding more knots to B-Spline function. In this paper, a self-adapted and multi-scale scheme for elimination of data redundancy is proposed by using many-knot spline fitting. And the validity of the algorithm is demonstrated by using it in different type of sampling data processing. The algorithm takes full advantage of Many-knot spline function’s advantages including cardinal type, explicit calculation (no need to solve system of equations) and localization, and can be use to reduce data redundancy with self-adapted mode, based on the idea of multi-scale analysis. The goal that uses a little amount of data to displace original large amount of data is achieved. It’s a novel idea to other similar schemes, and it can find application in data compression or pretreatment (or after-treatment) process of exiting data compression schemes.更多还原