Haar Wavelet..

Hey, I have to basically input a 128*128 Haar Wavelet Matrix. Thing is, I'll either have to code it in or hardcode the values myself. Does anyone know any inbuilt class or any function I can use to save any time????

Answer 1

[Wayne King]

Hi Aditya, Can you be more specific what you mean by the Haar wavelet matrix? Are you attempting to implement a 1-D DWT using the Haar wavelet? And if so, is it necessary that you implement as a matrix multiplication. If you have the Wavelet Toolbox, there are much more efficients ways to do it.

Wayne

[Aditya]

Hey! Thanks for your quick reply. I want a 2-D Haar Matrix...

like... 1 1 1 1

1 1 -1 -1

1 -1 0 0

0 0 1 -1

only well, for 128 rows and columns. I will then carry out the squaring of elements and division. Is there any inbuilt function or code which foes all this? Or will I have to do it on my own?

Answer 2

[mukala gayatri]

function H=haarmtx(n)

% HAARMTX Compute Haar orthogonal transform matrix.

%

% H = HAARMTX(N) returns the N-by-N HAAR transform matrix. H*A

% is the HAAR transformation of the columns of A and H'*A is the inverse

% transformation of the columns of A (when A is N-by-N).

% If A is square, the two-dimensional Haar transformation of A can be computed

% as H*A*H'. This computation is sometimes faster than using

% DCT2, especially if you are computing large number of small

% Haar transformation, because H needs to be determined only once.

%% Class Support

% -------------

% N is an integer scalar of class double. H is returned

% as a matrix of class double.

%

%

% I/O Spec

% N - input must be double

% D - output DCT transform matrix is double

%

% Author : Frederic Chanal ([e-mail]) - 2004

a=1/sqrt(n);

for i=1:n

H(1,i)=a;

end

for k=1:n-1

p=fix(log2(k));

q=k-2^p+1;

t1=n/2^p;

sup=fix(q*t1);

mid=fix(sup-t1/2);

inft=fix(sup-t1);

t2=2^(p/2)*a;

for j=1:inft

H(k+1,j)=0;

end

for j=inft+1:mid

H(k+1,j)=t2;

end

for j=mid+1:sup

H(k+1,j)=-t2;

end

for j=sup+1:n

H(k+1,j)=0;

end

end

[Laura Carballo-Sigler]

this is awesome!