Exporting Graphs from Matlab App in PNG or JPG

I'm using an app to fit some experimental data to model. The app wasn't written by myself. After fitting, few different graphs are shown with the results. I don't see the option to export these graphs in PNG or JPG or any other format for that matter inside the app window. Do you know how can one export graphs from this Matlab app.

Here is the app window shown. Usually, with graphs, there is an export button on the axes toolbar. However, none exist here.

[Walter Roberson]

ECRTOOLS by F. Ciucci appears to have been released in 2013, so it would be written in GUIDE rather than the more modern App Designer.

Unfortunately I do not find any source for it, so I cannot examine the code to estimate the difficulty of adding a save graph button

[Dario]

I do have a access to code and was able to found the lines where plotting is done. I tried using save function to save the data as .txt file and to plot it later outside Matlab. However, my text file remains empty after saving.

Here are lines of code and plot where I wanted to save the data.

Line 346 is what I added.

If you want, I can send you the code (zip file). There is a small installation procedure of the toolbox provided. Everything is explained in the manual.pdf

[DGM]

This works fine for me. If I press the "Draw ECR" button, I get a text file with three lines of data.

[Dario]

Update: I found ECR.txt file in the directory where the code is saved: OPTI-master folder. Thank you for trying this.

However, there is no additional .txt file in the same folder where I saved other graph on the right (confidence intervals for parameters).

Here is the line of code I added (line 457). It's supposed to save the confidence interval data in the separate file, conf.txt

[DGM]

The arguments passed to save() should be variable names, but boundary_cov(2,:) isn't a variable itself. This limitation makes more sense when you expect the output of save() to be a .mat file, wherein the specified variables are stored with their original names.

You could create new variable names, but since you're using save() with ascii mode output where the names are just discarded, these two row vectors are written on sequential lines in the output file exactly as if they were a 2D array. Without names, they're just two adjacent row vectors in the same order as they were originally.

boundary_cov_x = boundary_cov(1,:);
boundary_cov_y = boundary_cov(2,:);
save('conf.txt','boundary_cov_x','boundary_cov_y','-ascii')

Instead, you can just save boundary_cov as a whole 2D array. This produces the same result without the extra steps.

save('conf.txt','boundary_cov','-ascii')

[Dario]

Oh, okay. Basically, the whole variable can only be stored and not only part of it. Thank you for the help as it worked for me.

Answer 1

[ScottB]

One way to export to png is with the print command:

% Export to graphics file
savename = 'MyFile';
print(gcf,savename,'-dpng')

Answer 2

[Image Analyst]

Assuming it's the current/active figure, try

exportgraphics(gcf, 'ECRTools.png');

in the command window.

help exportgraphics

  EXPORTGRAPHICS  Save plot or graphics content to file
    EXPORTGRAPHICS(OBJ, FILESPEC) saves the specified graphics to a file.
    OBJ is the handle of any type of an axes, a figure, a chart that can 
    be a child of the figure, a tiled chart layout, or a container within 
    the figure. FILESPEC is a character vector or string scalar specifying
    the name of a file to create. It must include the extension, and can
    optionally include a full or relative path. EXPORTGRAPHICS supports 
    creating JPEG, PNG, and TIFF images, and EPS, PDF, and EMF vector
    formats (EMF is only supported on Windows).
 
    EXPORTGRAPHICS(___, OPTIONS) saves the graphics to the file, applying  
    the specified OPTIONS. OPTIONS is one or more of the following 
    name-value pairs.
 
    'ContentType', CONTENTTYPE  how to generate content in vector format files
    CONTENTTYPE can be one of the following values
    - 'vector' to generate vectorized (scalable) output 
    - 'image'  to generate rasterized output (embed image) 
    - 'auto'   allow MATLAB heuristic to decide
    The default value is 'auto'
 
    'Resolution', DPI  specifies output resolution for images. DPI must be
    a positive whole number. The default value is 150
  
    'BackgroundColor', COLORSPEC  specifies the color to use as the background 
    COLORSPEC can be one of the following values
    - a color name, such as 'red', or 'w'
    - an RGB triplet, such as [1 0 0] or [1 1 1]
    - a hexadecimal color code, such as '#FFF' or '#FFFFFF')
    - 'none' to use either a white or transparent background, depending on 
       what the output format supports 
    - 'current' to use the graphic's current background color
    The default value is 'white'
  
    'Colorspace', COLORSPACE  the color space to use when generating output
    COLORSPACE can be one of the following values
    - 'rgb'  generates RGB true color output
    - 'gray' generates grayscale output
    - 'cmyk' generates cyan, magenta, yellow, black output
    The default value is 'rgb'
 
    'Append', APPEND  specifies whether to append the output to an existing file or replace the content.
    APPEND can have a value of true or false.
    - true appends the output to an existing file
    - false replaces the content in an existing file
    The default value is false
 
    Example: Export a plot to a pdf file 
       plot(rand(3)); 
       exportgraphics(gca, 'results.pdf'); 
 
    Example: Export a plot to a pdf file and ensure output is vectorized/scalable 
       plot(rand(3)); 
       exportgraphics(gca, 'results.pdf', 'ContentType', 'vector');
 
    Example: Export a plot as an image to a pdf file 
       plot(rand(3)); 
       exportgraphics(gca, 'results.pdf', 'ContentType', 'image');
 
    Example: Export a plot and append to an existing pdf file
       plot(rand(3)); 
       exportgraphics(gca, 'results.pdf', 'Append', true);
 
    Example: Export a specific plot from the figure
       fig = figure; 
       ax1 = subplot(2,1,1); bar(magic(4));
       ax2 = subplot(2,1,2); plot(rand(4));
       exportgraphics(ax2, 'results.png'); 
 
    Example: Export all plots from figure to an image file at 300DPI
       subplot(2,1,1); bar(magic(4));
       subplot(2,1,2); plot(rand(4));
       exportgraphics(gcf, 'results.png', 'Resolution', 300); 
 
    Example: Export a plot to an image file with a red background 
       plot(rand(3)); 
       exportgraphics(gca, 'results.png', 'BackgroundColor', [1 0 0]);
  
    Example: Export a plot to a grayscale image file  
       plot(rand(3)); 
       exportgraphics(gca, 'results.png', 'Colorspace', 'gray');
  
    See also COPYGRAPHICS, EXPORTAPP

    Documentation for exportgraphics
       doc exportgraphics

Answer 3

[DGM]

Using gcf() won't work. The GUI handles aren't immediately visible, so calling gcf will just open and capture an empty figure.

This will capture the entire figure:

% find all the axes objects in the GUI figure
hf = findobj(findall(0),'type','figure','name','ECRTOOLS_GUI')

% save a screenshot of the figure
thistime = datestr(now,'yyyy-mm-dd_hh:MM:ss');
thisname = sprintf('capturedgui_%s.png',thistime);
exportgraphics(hf,thisname)

This will capture the three axes as separate images instead:

% find all the axes objects in the GUI figure
hf = findobj(findall(0),'type','figure','name','ECRTOOLS_GUI')
hax = findobj(hf,'type','axes')

% save each axes as a separate image
for k = 1:numel(hax)
    thistime = datestr(now,'yyyy-mm-dd_hh:MM:ss');
    thisname = sprintf('capturedaxes_%s_%02d.png',thistime,k);
    exportgraphics(hax(k),thisname)
end

Likewise, you could still fetch the plotted data externally. You just have to find it.

% find all axes objects in the GUI figure
hf = findobj(findall(0),'type','figure','name','ECRTOOLS_GUI');
hax = findobj(hf,'type','axes');

% sort through the axes to find the underlying data
% the main plot consists of three line objects in two overlaid axes
time = [];
sigma_n = [];
error_meas = [];
for k = 1:numel(hax)
    % look for the axes corresponding to sigma_n
    % this axes contains two line objects for sigma_n and sigma_n_meas
    if strcmp(hax(k).YLabel.String,'$\sigma_n$')
        hl = findobj(hax(k),'type','line');
        time = hl(2).XData;
        sigma_n = hl(2).YData; % hl(2) is sigma_n, not sigma_n_meas
    % look for the axes for the error measurement
    elseif strcmp(hax(k).YLabel.String,'$100\times(\sigma_n-\sigma_n^{\rm meas}~)$')
        hl = findobj(hax(k),'type','line');
        error_meas = hl.YData/100; % rescale back to unit-scale
    end
end

% save the data
save('ECR.txt','time','sigma_n','error_meas','-ascii')

Again, all of these examples are being external to the GUI code.