Cutting up image into smaller blocks - how to name resulting blocks according to grid pattern?

Hello All,

Total Matlab noob here. Thanks to these forums, I've managed to split a larger image into a grid of much smaller images. I've also managed to save each of these smaller images as a .png. However, I wish to save these images with file names relevant to their location in the grid. (The upper left block would be A_001, the block one to its right would be B_001, the one below that would be B_002, etc.) However, I am unsure how to do this. I'll show the relevant code below.

plotIndex = 1;
numPlotsR = size(ca, 1);
numPlotsC = size(ca, 2);
for r = 1 : numPlotsR
    for c = 1 : numPlotsC
        fprintf('plotindex = %d,   c=%d, r=%d\n', plotIndex, c, r);
        % Specify the location for display of the image.
        subplot(numPlotsR, numPlotsC, plotIndex);
        % Extract the numerical array out of the cell
        % just for tutorial purposes.
        rgbBlock = ca{r,c};
        imshow(rgbBlock); % Could call imshow(ca{r,c}) if you wanted to.
        [rowsB columnsB numberOfColorBandsB] = size(rgbBlock);
     
        drawnow;
        % Increment the subplot to the next location.
        plotIndex = plotIndex + 1;
    end
end
% Display the original image in the upper left.
subplot(4, 6, 1);
imshow(rgbImage);
title('Original Image');


for K = 1 : numel(ca)
    filename = sprintf('%d.png', K);
    imwrite(ca{K}, filename);
end

The first, much larger part of the code shows how we define %d. For the very first block, %d is 1, for the second, %d is 2, etc. The last few lines at the very end show how we name each individual photo file after %d. However, this is as far as I've gotten. I have tried altering the bottom lines, but everything I've tried has resulted in an error. Is there a way to name images in the grid fashion by using %d? Do I need a whole nother system all together?

Thank you!

Answer 1

[Image Analyst]

David:

Try this:

numPlotsR = 3;
numPlotsC = 37; % or size(ca, 2);
folder = pwd; % Or wherever you want.
for r = 1 : numPlotsR
    for c  = 1 : numPlotsC
	col = ExcelCol(c);
        baseFileName = sprintf('%s_%03d.png', col{1}, r);
	fullFileName = fullfile(folder, baseFileName);
	fprintf('Writing %s...\n', fullFileName);
%       imwrite(ca{r,c}, fullFileName);
    end
end

Looks like

...

Writing C:\Users\David\Matlab\work\Tests\AI_003.png...

Writing C:\Users\David\work\Tests\AJ_003.png...

Writing C:\Users\David\work\Tests\AK_003.png...

and so on.

ExcelCol.m is attached and converts numbers to the Excel column letter codes like you asked for.

[David Bukowski]

Thanks for replying! Just a few question, I'm pretty new at this. Do I replace my last few lines of code with this? Or do I paste this after it? Secondly, what do the numbers 3 and 37 mean? Should I change them? And thirdly, how does ExcelCol.m work? Should it be a separate file open in Matlab? Or do I copy and paste all of it into my existing file? And fourthly, what does "pwd" from "folder = pwd" mean?

Thank you for your patience!

[Image Analyst]

David:

No 3 and 37 were just some numbers I picked because I didn't have your actual cell array and I needed to pick something to get the code to run. You just use the values you already got, which I presume was from something like this:

numPlotsR = size(ca, 1);
numPlotsC = size(ca, 2);

replacing ca with the actual name you used for your cell array of course. Then use those dimensions in the loop below.

folder = pwd; % Or wherever you want.
for r = 1 : numPlotsR
    for c  = 1 : numPlotsC
	col = ExcelCol(c);
        baseFileName = sprintf('%s_%03d.png', col{1}, r);
	fullFileName = fullfile(folder, baseFileName);
	fprintf('Writing %s...\n', fullFileName);
        imwrite(ca{r,c}, fullFileName);
    end
end

folder=pwd just puts the "current folder" into a string called folder, which is then used in fullfile() to construct the full path name (folder + base file name + extension) of your image file.

ExcelCol is just some code I got off the File Exchange (I might have modified it though to improve it). It converts column numbers into letters like Excel column letters, like 27 = AA, 28=AB, etc. You can edit it if you want to see the details of how it works. Just put that file in your current folder, or in some other folder, like a "Utilities" folder that you have on the MATLAB path (set by the SetPath button on the Home tab of the tool ribbon).

[David Bukowski]

I think I understand most of that, thank you. However, when I run the code, I don't think I get individual files for each block like I used to. Any idea where these pictures are saving to?

Here's what I put in the code just now:

function Out=ExcelCol(In)

folder = pwd; % Or wherever you want.

numPlotsR = size(ca, 1);

numPlotsC = size(ca, 2);

for r = 1 : numPlotsR

for c = 1 : numPlotsC

col = ExcelCol(c);

baseFileName = sprintf('%s_%03d.png', col{1}, r);

fullFileName = fullfile(folder, baseFileName);

fprintf('Writing %s...\n', fullFileName);

imwrite(ca{r,c}, fullFileName);

end

end

end

[Image Analyst]

You should. They should be in the current folder -- same place where your script lives. If you want them somewhere else the specify it in folder:

folder = 'c:\my Images\this experiment'; % or wherever.

What filenames get printed out when you run it? That should say EXACTLY where they are being saved to.

[David Bukowski]

Hi there,

I am not seeing files show up anywhere. Should they appear in the Current Folder window on the left side? I believe the issue might actually be with the ExcelCol file? I'm getting an error message about 'In' when I run it, I'll show you the screenshot here.

[Image Analyst]

Upload your script where you're calling ExcelCol.

Answer 2

[Geoff Hayes]

David - ca seems to represent your cell array of blocks. The code you have (already) to write the blocks is

for K = 1 : numel(ca)
    filename = sprintf('%d.png', K);
    imwrite(ca{K}, filename);
end

can be modified to do (what I think you want) as

for r = 1 : numPlotsR
    for c  = 1 : numPlotsC
        filename = sprintf('%c_%03d.png', char(c + 64), r)
        imwrite(ca{r,c}, filename);
    end
end

In the above, we convert the integer columns (offset by 64) 65, 66, 67, ... to their ASCII equivalent of 'A', 'B', 'C", ...

[David Bukowski]

Thank you very much! Just one more question, how would you name the files if there are more than 26 columns? How would you name things AA_001, AA_002, etc? Because the images I want to cut up will eventually be very large. I tested it with this code, and after Z, it starts naming columns with [ and /, which gives me these errors.

Error using imwrite (line 548)

Unable to open file "\_001.png" for writing. You might not have write permission.

Error in attempt1multipleletter (line 89)

imwrite(ca{r,c}, filename);

[David Bukowski]

Alternatively, I think it would be fine if it started with AA rather than starting with A, B, C, etc before reaching AA after column Z.

Also, could you please explain why the integer columns are offset by 64?

[Geoff Hayes]

David - please see the ASCII tables from https://en.wikipedia.org/wiki/ASCII. Note that the integer 65 is mapped to the character 'A', 66 to 'B', etc. and so that is why we offset by 64. You could start with AA if you wish. The code would need to be updated to account for that.

[David Bukowski]

Thank you Geoff. If I wanted to make it AA, would the line of code just be:

filename = sprintf('%c_%03d.png', char(c + 64)char(c + 64), r) ?