Implementation of orthogonal wavelet transforms and their applications

In this paper the efficient implementation of different types of orthogonal wavelet transforms with respect to practical applications is discussed. Orthogonal single-wavelet transforms being based on one scaling function and one wavelet function are used for denosing of signals. Orthogonal multiwavelets are based on several scaling functions and several wavelets. Since they allow properties like regularity, orthogonality and symmetry being impossible in the single-wavelet case, multiwavelets are well suited bases for image compression applications. With respect to an efficient implementation of these orthogonal wavelet transforms approximating the exact rotation angles of the corresponding orthogonal wavelet lattice filters by using very few CORDIC-based elementary rotations reduces the number of shift and add operations significantly. The performance of the resulting, computationally cheap, approximated wavelet transforms with respect to practical applications is discussed in this paper.