Question cell structure, plotting and how to avoid a loop.

Hello everybody,

I have a 1x13 cell structure called "Im" containing 13 matrices of 512x512 dimension with integer number ( which represent images). I want to plot all 13 matrices along with a fit I made, which basically should be like this code:

    plot(fit,Pixels,Im{:}(:,256))

As I am sure you know, I get an error called Bad cell reference operation. The code works if I choose just one cell, for example plot(fit,Pixels,Im{5}(:,256)) but doesn't work as I tried above. I would wish not to use a loop as I think it could be avoided.

Thank you very much in advance

Answer 1

[Guillaume]

Even if Im{:}(:, 256) was valid matlab syntax it still wouldn't work as plot expects a pair of x and y coordinate for each line, not just one x vector and a number of y vectors.

The way to extract column 256 from all the images is with:

cellfun(@(img), img(:, 256), Im, 'UniformOutput', false) %return a cell array of columns

plot also accepts a vector of x and matrix of y where each column or row is a line to plot, so I would just transform your image columns in a matrix and concatenate that with Pixels. Assuming that Pixels is a column vector:

 plot(fit, [Pixels cell2mat(cellfun(@(img) img(:, 256), Im, 'uni', false))]);

edited for spelling and typos

[Filippos]

This looks very close to what I want.. I am however not very familiar with the cellfun command.. Right now it is giving me an error: 'Error: Expression or statement is incorrect--possibly unbalanced (, {, or [.'

What does the "img" variable mean? Do you maybe mean the "Im" variable?

[Image Analyst]

If you had uploaded your data he might have been able to try something and save you a lot of time. Oh well, there's still time to do it.

[Guillaume]

Sorry, there was an extra comma between @(img) and img(..) that shouldn't have been there (now corrected).

@(img) img(:, 256) is an anonymous function. It is equivalent to:

function out = myfun(img)
  out = img):, 256);
end

The whole cellfun is equivalent to:

 c = cell(size(Im));   
 for idx = 1:numel(Im)
   img = Im{idx};
   c{idx} = img(:, 256);
 end

[Filippos]

Many thanks, it has worked exactly as I wanted. Thank you very much and have a nice day sir.

Answer 2

[Image Analyst]

I don't know what fit and Pixels are but presumably they're lists of x and y coordinates. Then, for the next argument of plot() you don't put a whole image. You display images with a separate call to imshow(), image(), or imagesc().

% First, loop to display all 13 images.
for k = 1 : length(Im)
  thisImage = Im{k}; % Extract image
  subplot(3, 5, k);
  imshow(thisImage);
end
% Now plot fit vs. Pixels
subplot(3, 5, 15);
plot(fit, Pixels, 'b-', 'LineWidth', 3);
grid on;
xlabel('fit', 'FontSize', 20);
ylabel('Pixels', 'FontSize', 20);
title('Pixels vs. fit', 'FontSize', 20);

[Filippos]

I am sorry that I was not very clear. I do now want to produce the images, but the intensity profile in the middle of the image. The "fit" variable is the one you get from the cftool and the Pixels is just a vector 1:512. In the end I am planning to do something like this : http://tinypic.com/view.php?pic=2j0dnqr&s=8#.VB7KhFfitNg

[Guillaume]

Doesn't my answer below produce what you want?