<< wden denoising wnoise >>

swt >> denoising > wdencmp

wdencmp

De-noising or compression using wavelets.

Calling Sequence

[XC,CXC,LXC,PERF0,PERFL2] = wdencmp('gbl',X,'wname',N,THR,SORH,KEEPAPP)
[XC,CXC,LXC,PERF0,PERFL2] = wdencmp('gbl',C,L,W,N,THR,SORH,KEEPAPP)
[XC,CXC,LXC,PERF0,PERFL2] = wdencmp('lvd',X, 'wname',N,THR,SORH)
[XC,CXC,LXC,PERF0,PERFL2] = wdencmp('lvd',C,L, 'wname',N,THR,SORH)
[XC,CXC,LXC,PERF0,PERFL2] = wdencmp('lvd',X, 'wname',N,THR,SORH)
[XC,CXC,LXC,PERF0,PERFL2] = wdencmp('lvd',C,L, 'wname',N,THR,SORH)

Parameters

X :

input signal (1-D or 2-D)

C,L:

wavelet decomposition structure

THR :

positive threshold

XC:

de-noised or compressed version of X

[CXC,LXC]:

wavelet decomposition of XC

PERFL2 and PERF0:

are L^2 recovery and compression scores in percentages.

PERFL2:

PERFL2 = 100*(vector-norm of CXC/vector-norm of C)^2

N:

level of Wavelet decomposition

'wname':

is a string containing the wavelet name.

SORH('s' or 'h') :

soft or hard thresholding

KEEPAPP:

= 1, approximation coefficients cannot be thresholded, otherwise it is possible.

'gbl':

using one threshold value

'lvd':

level-dependent thresholds (THR must be of length N). For 2-D case THR must be a matrix of size 3 by N in the three orientations horizontal, diagonal and vertical.

Description

performs a de-noising or compression process of a signal or an image using wavelets.

Examples

x=sin(2*%pi*(0:0.01:1));
xn=x+rand(x);
thr=35;
//compression  using global thresholding
[xcomp,cxd,lxd,perf0,perfl2] = wdencmp('gbl',xn,'db3',3,thr,'h',1);
//denoising
// Find default values
[thr,sorh,keepapp] = ddencmp('den','wv',x);
// De-noise signal using global thresholding option.
xd = wdencmp('gbl',xn,'db3',4,thr,sorh,keepapp);
scf();clf();
plot(xn);
plot(xcomp,'r');
plot(xd,'g');
legend(["noisy signal","compressed signal","de-noised signal"],1);

See also

Authors


Report an issue
<< wden denoising wnoise >>