How to obtain a curvelet transform matrix?

I want to obtain a matrix C in matlab which is the n-by-n DCT (discrete curvelet transform) matrix such that for a given set of signals X and given set of coefficients A (I supposedly think which will better represent edges of X ) we can get a representation X=C*A. C will be a universal transform matrix like an n-Haar transform matrix. Can I obtain such a matrix since curvelet is linear? I have curvelab 2.1.3 installed and a function fdct_usfft.m returns curvelet transform of a given input. But I need the curvelet transform to be a transform operator matrix rather than as an operator on a signal.

[晓珍 李]

Hello, may I share your code after adjusting the parameters?I also want to get the Qubo transform matrix

Answer 1

[Matt J]

This FEX submission might help you,

http://www.mathworks.com/matlabcentral/fileexchange/44669-func2mat--convert-linear-function-to-matrix

[Mrin San]

Hi thanks for replying. In the function func2mat(fun,Xtypical,varargin), how do we have to give input fun? Is it like giving access to the matlab function (in my case curvelet function)?

[Matt J]

It can be any MATLAB function handle, like an anonymous function or a handle to an mfile. In your case, it would probably be something like this

C=func2mat(@fdct_usfft, one(n,1))

[Mrin San]

Hi I have tried to use the code using the syntax

 A=func2mat(@fdct_usfft,x)

where x is a 512*512 array of real data or complex data (i checked both). Matlab throws an error as following -------

Undefined function 'find' for input arguments of type 'cell'.

Error in func2mat (line 79)

             [I{j}, ~, S{j}] = find(T(:));

Can you please throw some light where i went wrong?

[Matt J]

what is the output of

 >> class(fdct_usfft(x))

[Mrin San]

Also I computed a simple discreet cosine transform operator matrix using

 dctmtx(4)

command and computed a transform matrix using a sample array

x= [1 2; 2 4];

Both are different. Will you please explain why they can be different?

[Mrin San]

The output of class(fdct_usfft(x))is cell

[Matt J]

The output of the function you pass to func2mat must be numeric, just as if you had multiplied by the matrix.

[Mrin San]

Thanks a lot Sir. Trying to tweak its parameters out to get the right matrix out of it.

[Matt J]

where x is a 512*512 array of real data or complex data (i checked both)

Note as well that if the matrix is intended to represent a 2D separable transform, it can be more efficient to compute the 1D transform matrix instead, and combine them using this FEX submission.

See here for a related example.