📄 waveshrink.m
字号:
function [xh, wcoef] = WaveShrink(y,type,L,qmf)
% WaveShrink -- Soft Threshold Shrinkage Applied to Wavelet Coefficients
% Usage
% [xh,xwh] = WaveShrink(y,type,L,qmf)
% Inputs
% y 1-d signal. length(y)= 2^J
% Normalized to noise level 1! (See NoiseNorm)
% type string. Type of shrinkage applied:
% 'Visu','SURE','Hybrid','MinMax','MAD'
% Optional; default == 'Visu'
% L Low-Frequency cutoff for shrinkage (e.g. L=4)
% Should have L << J!
% qmf Quadrature Mirror Filter for Wavelet Transform
% Optional, Default = Symmlet 8.
% Outputs
% xh estimate, obtained by applying soft thresholding on
% wavelet coefficients
% xwh Wavelet Transform of estimate
%
% Description
% WaveShrink smooths noisy data presumed to have noise level 1
% by transforming it into the wavelet domain, applying soft
% thresholding to the wavelet coefficients and inverse transforming.
%
% The theory underlying these methods is described in a variety of
% papers by D.L. Donoho and I.M. Johnstone.
%
% The different methods of selecting thresholds are detailed
% in their articles.
%
% See Also
% FWT_PO, IWT_PO, MakeONFilter, NoiseNorm, RigorShrink
%
if nargin < 2,
type = 'Visu';
end
if nargin < 3,
L = 3;
end
if nargin < 4,
qmf = MakeONFilter('Symmlet',8);
end
%
[n,J] = dyadlength(y) ;
wcoef = FWT_PO(y,L,qmf) ;
%
if strcmp(type,'Visu'),
wcoef((2^(L)+1):n) = VisuThresh(wcoef((2^(L)+1):n)) ;
elseif strcmp(type,'SURE'),
wcoef = MultiSURE(wcoef,L);
elseif strcmp(type,'Hybrid'),
wcoef = MultiHybrid(wcoef,L);
elseif strcmp(type,'MinMax'),
wcoef((2^(L)+1):n) = MinMaxThresh(wcoef((2^(L)+1):n)) ;
elseif strcmp(type,'MAD'),
wcoef = MultiMAD(wcoef,L);
end
%
xh = IWT_PO(wcoef,L,qmf);
%
% Copyright (c) 1993-5. Jonathan Buckheit, David Donoho and Iain Johnstone
%
%
% Part of WaveLab Version 802
% Built Sunday, October 3, 1999 8:52:27 AM
% This is Copyrighted Material
% For Copying permissions see COPYING.m
% Comments? e-mail wavelab@stat.stanford.edu
%
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -