查看文章

Wavelets transform, particularly the Discrete Wavelet Transform (DWT) is an important problem for many applications. From the mathematical point of view many different wavelet transforms are developed, out of which orthogonal wavelet transforms are more important. Various algorithms for their computation are developed as well. This paper describes parallel implementation of the 2D Haar and Daubechies D4 wavelet transform in mesh architecture of processes. At the beginning, a short mathematical introduction is presented. After that the parallel algorithms implementations are analyzed. In this paper is shown that the mathematical complexity of those algorithms is the same. The complexity is)(log2 n O in case of one point in one process and)/( 2 pnO in case of more points in one process. The practical implementation is realized as well. For practical implementation are used the C++ programming language and …