Mean of a bunch of slices

% -----------------Step 3 -Masking the PCASL Files
PCASL_Files_Masked= PCASL_Files.*Human_Mask;          %apply the mask. PCASL Files Masked = 64x64x128x16, Human_Mask=64x64x128x16

PCASL_Files_Masked=mean(PCASL_Files_Masked,4);        %Taking the mean of each image, now PCASL_Files_Masked= 64x64x128

PCASL_Files_Masked=double(PCASL_Files_Masked);          
save('PCASL_Files_Masked','PCASL_Files_Masked');        %Saving the PCASL_Files_Masked

Question 1) When I am taking the mean of all 16 slices I do PCASL_Files_Masked=mean(PCASL_Files_Masked,4); Is this theoretically correct? I am just not sure about using the '4'

Question 2) Next I want to get the mean over each image, however only in the 'masked' area

PCASL_Files_Masked_1= mean(mean(PCASL_Files_Masked));   % gives the mean signal over each measurement or each brain volume (there's 128 of them), so mean of (64, 64, 1:128)

But it doesn't do the job correctly. I want to ONLY take the mean over where the BRAIN is, not the rest of the space (which is zero) Just to clarify again, I believe when Matlab takes the mean over the image it even accounts for the number of pixels in the zero area(outside of the brain)

As in :

mean=[brain signal(which is some value)+ outside brain(which is zero)]/ [number of pixels in brain area+ number of pixels outside of brain] which is theoretically wrong.

Thanks in advance

Answer 1

[Jan]

Q1. mean(X, 4) calculates the mean over the 4th dimension. Therefore I assume your command is correct. I suggest to try it using a tiny test matrix.

Q2. If you want to ignore the zeros in a mean, simply use sum(X) and divide by sum(X ~= 0).

[Image Analyst]

I think there are only 3 dimensions - I think the final x16 meant 16 bit integers, not an additional dimension.

[Med_Imager]

There are 4 dimensions at first

64x64x128x16

size of each image=64x64

# of repetitions=128

Slices in each image=16

Answer 2

[Image Analyst]

If you have a mask image you can do

pixelsInsideMask = sliceImage(maskImage); % This is a 1D vector of pixels only within the mask area.
meanInsideMask = mean(pixelsInsideMask);

Now sliceImage and even maskImage can vary as you move through the volume slice by slice. To keep track of meanInsideMask as a function of slice, make it an array

for slice = 1 : totalNumberOfSlices
  % First somehow get sliceImage.
  % Next somehow get maskImage - the mask for this slice level.
  % Then, once you have those....
  pixelsInsideMask = sliceImage(maskImage); % This is a 1D vector of pixels only within the mask area.
  meanInsideMask(slice) = mean(pixelsInsideMask);
end