How do I create a matrix that has as values ​​the averages of variables found in loops?

My program analyzes a reference image: named '21', then cut out two ROIs from this image and save their coordinates. Then to the remaining images cut out the same number of ROIs at the same coordinates.

At the end of the program I will have 2*21=42 ROIs.

srcFile = dir('C:\Users\agnes\Pictures\pp4 tgc-med rd20\*.dcm');
pathname = ('C:\Users\agnes\Pictures\pp4 tgc-med rd20\');

% define these image
numberofimages = 21; 
numberofroi = 2;
% show the reference image
I=dicomread('21');
imshow(I)
% use imcrop to get rectangle coordinates
R = zeros(numberofroi,4);
for nr = 1:numberofroi
    [~,R(nr,:)] = imcrop(gca);
end
% crop corresponding ROIs from each non-reference image
roibasket = cell(numberofroi,numberofimages);
for ni = 1:numberofimages
    for nr = 1:numberofroi
        filename=(num2str(ni));       
        
        pileofimages=dicomread(strcat(pathname,filename));
        info=dicominfo(strcat(pathname,filename));
        
        roibasket{nr,ni} = imcrop(pileofimages,R(nr,:));
    end
end
% show the cropped image regions for demonstration
montage(roibasket.','size',[numberofroi numberofimages])

For these ROIs I have to compute the mean and save the result in a 21*2 matrix, who can I do that?

Answer 1

[Matt J]

If that's all you want to do, you shouldn't be saving the coordinates of the ROIs. You should just create a binary mask BW of each ROI and stack your images in a 3D stack A and do

sum(A.*BW,[1,2])./sum(BW,[1,2]);

[HelpAStudent]

Is not just what I want to do. I need first to do all I wrot in the code. And after with my 2*21 ROIs I nead to create a matrix wicht have for each value the mean of the each ROI.

[Matt J]

Better to use drawrectangle then.

R = nan(numberofroi,4);
masks=cell(numberofroi,1);%<---ADDED
for nr = 1:numberofroi
    h = drawrectangle(gca); wait(h); %CHANGED double-click roi to accept
    R(nr,:)=h.Position;%<---ADDED
    masks{nr}=h.createMask;%<---ADDED
end

roibasket = cell(numberofroi,numberofimages);
meanbasket = nan(size(roibasket));  %<---ADDED
for ni = 1:numberofimages
    for nr = 1:numberofroi
        filename=(num2str(ni));       
        
        pileofimages=dicomread(strcat(pathname,filename));
        info=dicominfo(strcat(pathname,filename));
        
        roibasket{nr,ni} = imcrop(pileofimages,R(nr,:));
        meanbasket(nr,ni) = mean( pileofImages(masks{nr})  );    %<---ADDED
        
    end
end

[HelpAStudent]

So, meanbasket should be the matrix in question right? Can I ask you why did you change imcrop with drawrectangle?

[Matt J]

Yes. Drawrectangle provides a more direct way to get both the roi box boundaries and a binary mask for it.

[HelpAStudent]

Thank you very much for helping. From the obtained matrix is ​​there a way to know when the values ​​repeat?

For example, if in the first column and first row I have the value x, when does it appear in another position?