改进的基于提升格式的DWT硬件实施方案

提升结构与传统的Mallat算法相比,降低了计算的复杂度,也提高了计算速度.最近又提出了一种翻转结构,它在提升格式的基础上进一步提高了计算速度.结合翻转结构,以9/7小波为例,设计了一种具有递归结构的硬件实施方案,实现了2级一维离散小波变换.并研究了翻转结构的系数量化方法,在整个设计中,实现了无乘法运算.测试结果证明,提出的这种硬件结构通过采用递归结构,把第二级小波变换插入到第一级变换的缝隙中,其硬件占用少,计算速度快,耗电量低.

Implementation of an Improved Hardware for DWT of Lifting Scheme

Lifting scheme requires less computation than conventional Mallat's algorithm.A new method called flipping architecture has been proposed recently,which speeds up further the computation for lifting scheme.A recursive architecture is thus designed on the basis of flipping architecture to implement the 2-level 1-D DWT of 9/7 wavelet,which will insert the 2nd-level DWT into the gaps during 1st-level DWT.Furthermore,the coefficient quantization of flipping architecture is also discussed aiming at eliminating multiplication in the whole design process.The experimental results show that the proposed architecture has many merits,such as faster speed,lower power consumption and hardware saving.