set opacity to M x N x P grayscale matrix

I have a matrix define as data = rand(M,N,P) where M = 262, N = 359, P = 562. Each element in the matrix gets assigned an intensity value for 0 - 255 depending on input scan data.

E.g. data(200,220,232) = 32 or data(200,220,233) = 156

How can i assign transparency to each element in the matrix based on the intensity value. For example: all intensities below 128 have 0.8 transparency, and every value between 128-255 has linear opacity from 0.1-0.4

This is all in Matlab.

Thank you

Answer 1

[Youssef Khmou]

Logical operations can divide the elements, here is the starting point :

 M=rand(262,359,562);
 Threshold=0.45; % an example

 E=M ;%  matrix of transparency
 E(E<Threshold)=0.8;
 E(E>Threshold)=0.4;

[varun]

So if I understand the solution correctly, I should have two matrices. One with transparency and the other with actual intensities between 0-255.

Does this mean, when I place this a 3D grid. I should overlay the two matrices?

Right now i have a matrix mapped_3d(282,359,562) that gets passed to

vol3d('Cdata',mapped_3d);view(3);

function and it generates this figure.

However, if I want to make all dark points in the volume transparent, should I just pass the transparency matrix to vol3d function?

[Youssef Khmou]

I can not understand your approach, the proposition i posted is from mathematical viewpoint, but making the transparency in 3D plot is something that needs more details

Answer 2

[Walter Roberson]

Transparency is a property of images, not of matrices. And there is no build-in method in MATLAB to create a 4-dimensional plot (three axis plus value at each point = 4 dimensions.) You have to define how you are planning to display the data, and then we can talk about transparency.

[varun]

I suspected that was the case. To visualize the matrix I am using vol3d function found at: http://www.mathworks.com/matlabcentral/fileexchange/22940-vol3d-v2

An example of how the data is shown here in this image. (Same as the one i answered in another answer).

It is generated by this line of code.

vol3d('cdata',mapped_3d);

where mapped_3d is the 3 dimensional matrix

[Walter Roberson]

I would have to look more closely at the program some time. I think it is creating patch objects with 3D volume areas (marching cubes algorithm I seem to recall.) If it is, then a patch() would have been generated; try

findobj(gca, 'type', 'patch')

Once you find that patch object, you can adjust it's AlphaCData property