Sum of a cell of arrays

My code consists of scanning the current directory for all .bmp image files, creating a cell array (called "C") with the names of these files, and then imread-ing them. This creates a cell of the array images.

My question is how to take the sum of each image array read to create a single new array? The sum function obviously doesn't work since it produces a single value output. Manually typing C{1} + C{2} (and so on) will get me the correct results, but I need this to be automated and to work for any number of image files. I have tried to do a recursive function as well, but apparently it doesn't like adding arrays together. My code is shown below:

files = dir(fullfile(pwd, '*.bmp'));
    C = cell(length(files), 1);
    for k = 1:length(files);
        names = files(k).name;
        C{k} = imread(names);
    end
        le = length(files);
        Final = recursive(le);

function r = recursive(x)
if (x == 0)
    total = 0;
else
    files = dir(fullfile(pwd, '*.bmp'));
    C = cell(length(files), 1);
    for k = 1:length(files);
        names = files(k).name;
        C{k} = imread(names);
    total = C{x} + recursive(x-1);
    end
end

r = total;

I of course have the first set of scripts under another function, but have not pasted it here. Any help would be appreciated.

Answer 1

[José-Luis]

%dummy data
C = cell(10,1);
m = 15; %image size
n = 18;
numImages = 20;
for ii = 1:numImages
    C(ii) = {rand(m,n)};
end
%The sum:
your_sum = sum ( reshape (cell2mat(C),m,n,numImages), 3 );

[Michael]

Thanks for replying José-Luis!

I've tried adapting your code towards mine and came up with this:

files = dir(fullfile(pwd, '*.bmp'));
    C = cell(length(files), 1);
    for k = 1:length(files);
        names = files(k).name;
        C{k} = imread(names);
        si = size(C{1});
        m = si(1);
        n = si(2);
        le = length(files);
        your_sum = sum( reshape (cell2mat(C),m,n, le), 3);
    end

However, I realized that my images files have a 3rd dimension for some reason and are of type uint8. With my sample data, my cell array is of this format:

C =

    [541x541x3 uint8]
    [541x541x3 uint8]
    [541x541x3 uint8]
    [541x541x3 uint8]

Running the code while disregarding the third dimension just gave me

??? Error using ==> cell2mat at 46
All contents of the input cell array must be of the same data type.

Do I need to account for the 3rd dimension in some way?

[José-Luis]

What does the sum of the images represent? Say you have two rgb vectors [0 255 127] and [255 0 255]. What should you get for the sum?

EDIT Also, I assumed that all your images are the same size. If not, you have to rescale them before you can add them.

[Michael]

Yes all the images will be the same size. The images I am combining are black and white images. Each image has different areas of white (or gray), and I am trying to see the total area of "whiteness" that occurs in the final composite picture. So for each pixel in the image, I am trying to find the total sum of its "whiteness" from the other pictures. A spot that is lit up in every picture should produce a bright white pixel, while other pixels that do not light up in every image should be gray in the final composite.

Therefore, the maximum sum of a pixel should result in a bright white pixel. I hope this helped answer your question.

[José-Luis]

And how would you define "whiteness" mathematically. Maybe you need to make your image grayscale first and pick the ones with the lowest sum.

Answer 2

[Michael sch]

so what you can do get length(x) and then loop for all x and then sum them k=1:length(x) y=y+x{k} end

   x{1}=[1 2; 3 4]
  x{2}=[ 5 6 ;7 8]
  x{1}+x{2}

ans =

     6     8
    10    12
length(x)

ans =

     2