摘要
文章提出了一种基于JPEG2000标准中无损压缩滤波器组的VLSI实现方法。在文章中我们充分利用了小波系数是2的幂的和的形式的性质、FIR滤波器的线性性质以及流水线的设计技术,不需要乘法器就完成了小波变换。本文所提出的设计方法的特征:没有为小波变换专门设计下采样模块,而是通过一个时钟分频方案来完成;通过移位加来代替乘法运算;无“等待时间”的开销;吞吐量高、芯片面积小。实验结果表明本文提出的方法是一种适合于VLSI实现的设计方案。
This paper proposes an efficient VLSI method for realizing Discrete Wavelet Transformation for lossless compression in the upcoming JPEG2000standard.Th e proposed method takes advantage of power-of-two format of wavelet coefficie nts;the linear phase property of FIR filters and pipeline technology to obtain multiplierless wavelet transformation.The main features of the proposed approac h are:No explicit down-sampling block is used.Down-sampling is effected usi ng a multi-clock scheme.Multiplierless implementation through performing shift-add operations.A″wait time ″free architecture.Comparing with the existing designs,this approach supports fast transformation and allows achieving very c ompact design.The results show that the proposed method is suitable for VLSI i mplementation.
出处
《微电子学与计算机》
CSCD
北大核心
2003年第3期5-7,共3页
Microelectronics & Computer
基金
国家创新研究群体科学资助项目(600024301)
