FWT2 - Fast Wavelet Transform 2D

Usage

c = fwt2(f,w,J);
c = fwt2(f,w,J,...);

Input parameters

Output parameters

Description

c=fwt2(f,w,J) returns wavelet coefficients c of the input matrix f using J iterations of the basic wavelet filter bank defined by w. Please see fwt for description of w and J.

fwt2 supports just the non-expansive boundary condition 'per' and critically subsampled filter banks in order to be able to pack the coefficients in a matrix. Also the J is limited to some maximum value for the same reason.

Additional flags make it possible to specify how the algorithm should subdivide the matrix:

'standard'

Standard behaviour of the JPEG 2000 standard. This is the default.

'tensor'

This corresponds to doing a full fwt along each dimension of the matrix.

Examples:

Some simple example of calling the fwt2 function, compare with the cameraman image. Only the 70 dB largest coefficients are shown, to make the structures more visible.

The first example uses the standard layout:

c = fwt2(cameraman,'db8',4);
imagesc(dynlimit(20*log10(abs(c)),70));
axis('image'); colormap(gray);

The second example uses the tensor product layout:

c = fwt2(cameraman,'db8',4,'tensor');
imagesc(dynlimit(20*log10(abs(c)),70));
axis('image'); colormap(gray);