perform_wavelet_transform.m

小波提升算法的实现

function y = perform_wavelet_transform(x, wavelet_type, VM, Jmin, dir, driver)

% perform_wavelet_transform - just a wrapper
%   to call the correct wavelet transform of WAVELAB or MATLAB (wavelet toolbox)
%   with the correct set of parameters, 
%   either in 1D or in 2D.
%
%   y = perform_wavelet_transform(x, wavelet_type, VM, Jmin, dir, driver);
%
%   'x' is either a 1D or a 2D array.
%   'wavelet_type' can be 
%       'daubechies', 'symmlet', 'battle', 'biorthogonal'.
%   'VM' is the number of Vanishing moment (both for primal and dual).
%   'Jmin' is the minimum scale (i.e. the coarse channel is of size 2^Jmin
%       in 1D).
%   'dir' is +1 for fwd transform and -1 for bwd.
%   'driver' can be either 'wavelab' or 'matlab'.
%
%   'y' is an array of the same size as 'x'. This means that for the 2D
%   we are stuck to the wavelab coding style, i.e. the result
%   of each transform is an array organized using Mallat's ordering
%   (whereas Matlab official toolbox use a 1D ordering for the 2D transform).
%
%   Here the transform automaticaly select symmetric boundary condition
%   if you use a symmetric filter. If your filter is not symmetric
%   (e.g. Dauechies filters) then as the output must have same length
%   as the input, the boundary condition are automatically set to periodic.
%
%   Copyright (c) 2004 Gabriel Peyr