Apply a mask to 4D medical images in a vectorized way

Dear All,

I have a 4D array V, with the size [144 144 12 25]. In fact, it is a 4D medical image with 12 stacked slices in 25 different time. I want to apply a mask, named M, to each image in the array in a vectorized way. M is a 144x144 array containing logical values. Could someone help me how to do it?

Thanks,

Ahmad

[José-Luis]

I notice that you don't seem accept many answers after asking questions. Some of the answers seemed valid to me. Please remember to accept answers if they help you, since it is the only "payment" the volunteers in this forum receive.

Answer 1

[José-Luis]

 data = rand(144,144,12,25);
 your_mask = rand(144,144) > 0.5;

Assuming you want to multiply by the mask:

 your_result = bsxfun(@times,data,mask);

Note that you can do operations other than multiplication. Please have a look at

 doc bsxfun

[AP]

Isn't bsxfun an elementwise operation?

[Sean de Wolski]

Isn't that what you want?

+1

[AP]

I want the operation on each image. I thought bsxfun is only for each pixel and can't be applied to each image in a 4D array. Please correct me if I am wrong: so when we call bsxfun(@times,data,mask), bsxfun applies the mask to each 144x144. Is that right?

[José-Luis]

Yes. That's what the "singleton expansion" in the documentation means.

Answer 2

[Conrad]

Something like this:

fV = V.*M(:,:,ones(1,size(V,3)),ones(1,size(V,4)));

Conrad

[AP]

How can I convert M, with size [144 144] to M with [144 144 12 25]?

[AP]

I think, I have to use

repmat(M, [1 1 size(V,3) size(V,4)])

to duplicate M in other dimensions and make it [144 144 12 25].