apply 3D DCT on a 3d matrix

Hello,

Is there any way to apply 3d dct on a 3d matrix without using the multidimensional function mirt_dctn?

i.e. working seperately on each dimension of the matrix.

Thank you!

Answer 1

[Shlok]

Hi Konstantina,

Yes, you can apply a 3D Discrete Cosine Transform (DCT) to a 3D matrix without using “mirt_dctn”. This can be done by applying a 1D DCT sequentially along each dimension. To achieve this, use the “dct” function present in Signal Processing Toolbox. Here’s an example of how to apply a 3D DCT manually:

A = rand(4, 4, 4); % Assuming a sample 3D matrix 

% Applying 1D DCT along each dimension separately 
A_manual = dct(A, [], 1); 
A_manual = dct(A_manual, [], 2); 
A_manual = dct(A_manual, [], 3); 

disp('Manual 3D DCT:'); 

Manual 3D DCT:

disp(A_manual); 

(:,:,1) =

    4.0438   -0.2585    0.0015    0.4368
   -0.2388    0.0352    0.3830    0.1899
    0.2355   -0.0452   -0.2289    0.0318
    0.2921    0.0377    0.4991    0.2263


(:,:,2) =

   -0.1791   -0.1629   -0.3286    0.4020
    0.2812    0.1543    0.2249   -0.1348
   -0.0387    0.0317    0.1924   -0.0612
    0.0865   -0.2584    0.3129   -0.2798


(:,:,3) =

    0.2543   -0.3150   -0.0084    0.1145
   -0.4315   -0.4029   -0.1167   -0.0776
   -0.1673   -0.0063    0.0381    0.2420
    0.4992    0.2309   -0.5001    0.1140


(:,:,4) =

    0.0050    0.0784   -0.2937    0.1079
    0.5732    0.2763   -0.0944   -0.2212
    0.1077    0.1326   -0.2218    0.2908
   -0.0236    0.2336    0.1926   -0.0552

To know more about “dct”, refer to the following MathWorks documentation link:

https://www.mathworks.com/help/signal/ref/dct.html