How can I load a .mat struct?

Hello community:

I use Matlab 2021a and when I try to load a .mat struct with these sentences load('linesStruct_1.mat') or load('linesStruct_1') I receive the following message:

Error using load

Unable to read MAT-file C:\ ... linesStruct_1.mat. Not a binary MAT-file. Try load -ASCII to

read as text.

Error in drawpoint_Algoritmo_Rename_Struct_Pendiente (line 5)

load('linesStruct_1.mat');

Could it be because I have saved the structures with a different name (lines in place of linesStruct_1.mat) than the image? Or maybe it has to do with the shortcut icons that appear in my folders?.

I remember this was not happening in other versions. I have been reading the many cases in which this occurs but I can not find a solution according to my situation. Any idea?

Thank you very much.

Answer 1

[Walter Roberson]

The file is not a .mat file. At the moment we do not know what it is. Experiment with

filename = 'linesStruct_1.mat';

if isunix()
  filename = 'handel.mat';
end

[fid, msg] = fopen(filename, 'r');
if fid < 0
   error('could not open file "%s" because "%s"', filename, msg);
end

bytes = fread(fid, [1 100], '*uint8');
fclose(fid)

ans = 0

fprintf('first part of file as text:\n');

first part of file as text:

fprintf('%s\n', char(bytes));

MATLAB 5.0 MAT-file, Platform: GLNX86, Created on: Tue Mar  9 11:01:23 2004                         

fprintf('\nas decimal: \n');

as decimal: 

disp(bytes);

  Columns 1 through 33

    77    65    84    76    65    66    32    53    46    48    32    77    65    84    45   102   105   108   101    44    32    80   108    97   116   102   111   114   109    58    32    71    76

  Columns 34 through 66

    78    88    56    54    44    32    67   114   101    97   116   101   100    32   111   110    58    32    84   117   101    32    77    97   114    32    32    57    32    49    49    58    48

  Columns 67 through 99

    49    58    50    51    32    50    48    48    52    32    32    32    32    32    32    32    32    32    32    32    32    32    32    32    32    32    32    32    32    32    32    32    32

  Column 100

    32

fprintf('\nas hex: \n');

as hex: 

fprintf('%02x ', bytes);

4d 41 54 4c 41 42 20 35 2e 30 20 4d 41 54 2d 66 69 6c 65 2c 20 50 6c 61 74 66 6f 72 6d 3a 20 47 4c 4e 58 38 36 2c 20 43 72 65 61 74 65 64 20 6f 6e 3a 20 54 75 65 20 4d 61 72 20 20 39 20 31 31 3a 30 31 3a 32 33 20 32 30 30 34 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 

fprintf('\n');

The bit about isunix() is my substituting a different file for the purpose of demonstration here.

You are hoping to see the text version start with 'MATLAB 5.0' or 'MATLAB 7.3' . If you show us the three parts of the output we might be able to tell you what kind of file you actually have instead of a .mat file.

[David Ruiz Velasco]

Hello again.

I show you the results obtained. I include the code with I saved the files. I apologize for not respecting the correct setting of the live script.

Name = sprintf("linesStruct_%i.mat", i);

Thanks in advance.

filename = 'linesStruct_1.mat';
if isunix()
  filename = 'handel.mat';
end
[fid, msg] = fopen(filename, 'r');
if fid < 0
   error('could not open file "%s" because "%s"', filename, msg);
end
bytes = fread(fid, [1 100], '*uint8');
fclose(fid)

ans = 0

fprintf('first part of file as text:\n');
first part of file as text:
fprintf('%s\n', char(bytes));

ÿØÿà JFIF ÿÛ

$.' ",#(7),01444'9=82<.342ÿÀ

fprintf('\nas decimal: \n');

as decimal:

disp(bytes);

Columns 1 through 32

255 216 255 224 0 16 74 70 73 70 0 1 1 0 0 1 0 1 0 0 255 219

0 67 0 8 6 6 7 6 5 8

Columns 33 through 64

7 7 7 9 9 8 10 12 20 13 12 11 11 12 25 18 19 15 20 29 26 31 30

29 26 28 28 32 36 46 39 32

Columns 65 through 96

34 44 35 28 28 40 55 41 44 48 49 52 52 52 31 39 57 61 56 50 60 46

51 52 50 255 192 0 11 8 1 0

Columns 97 through 100

1 0 1 1

fprintf('\nas hex: \n'); 

as hex:

fprintf('%02x ', bytes);

ff d8 ff e0 00 10 4a 46 49 46 00 01 01 00 00 01 00 01 00 00 ff db 00 43 00 08 06 06 07 06 05 08 07 07 07 09 09 08 0a 0c 14 0d 0c 0b 0b 0c 19 12 13 0f 14 1d 1a 1f 1e 1d 1a 1c 1c 20 24 2e 27 20 22 2c 23 1c 1c 28 37 29 2c 30 31 34 34 34 1f 27 39 3d 38 32 3c 2e 33 34 32 ff c0 00 0b 08 01 00 01 00 01 01

 fprintf('\n');

[Walter Roberson]

That 'JFIF' tells me that you somehow saved a JPEG file over top of your .mat file. And indeed https://www.ntfs.com/jpeg-signature-format.htm says that JPEG start with hex FF D8 FF, which your file does.

[David Ruiz Velasco]

Hello.

I think I have found the solution in this post:

https://es.mathworks.com/matlabcentral/answers/177752-how-do-i-open-my-matlab-figures.

I have been reading about the difference between images and figures. If I take account the used code, I should use saveas or savefig only after modifying the code. However, I understand that keyboard command will still be a problem.

I will keep working on this.

Thank you very much.

clear;
clc;

c=cell(10,1);

for i=1:1
    
    c{i}=imread(sprintf('Mexico_Resize_%d.jpg',i));
    columns = 256; rows = 256;
    
    I1 = c{i};
    imshow(I1);
    
    hold on
    I1 = imresize(I1, [columns rows]);
    
    x1 = 64; y1 = 64; x2 = 64; y2 = 128; x3 = 64; y3 = 192;
    x4 = 192; y4 = 64; x5 = 192; y5 = 128; x6 = 192; y6 = 192;
    
    roi1= drawpoint('Position', [x1, y1], 'Color','b');
    event.listener(roi1, 'ROIMoved', @allevents_1);
    roi2= drawpoint('Position', [x2, y2], 'Color','g');
    event.listener(roi2, 'ROIMoved', @allevents_2);
    roi3= drawpoint('Position', [x3, y3], 'Color','w');
    event.listener(roi3, 'ROIMoved', @allevents_3);
    roi4= drawpoint('Position', [x4, y4], 'Color','c');
    event.listener(roi4, 'ROIMoved', @allevents_4);
    roi5= drawpoint('Position', [x5, y5], 'Color','r');
    event.listener(roi5, 'ROIMoved', @allevents_5);
    roi6= drawpoint('Position', [x6, y6], 'Color','m');
    event.listener(roi6, 'ROIMoved', @allevents_6);
    
    % Stop at each iteration to move the points:  
    keyboard;
    
    dataA(1,1) = roi1.Position(1,1); dataA(1,2) = roi1.Position(1,2); 
    dataA(2,1) = roi2.Position(1,1); dataA(2,2) = roi2.Position(1,2);
    dataB(1,1) = roi2.Position(1,1); dataB(1,2) = roi2.Position(1,2); 
    dataB(2,1) = roi3.Position(1,1); dataB(2,2) = roi3.Position(1,2);
    dataC(1,1) = roi4.Position(1,1); dataC(1,2) = roi4.Position(1,2); 
    dataC(2,1) = roi5.Position(1,1); dataC(2,2) = roi5.Position(1,2);
    dataD(1,1) = roi5.Position(1,1); dataD(1,2) = roi5.Position(1,2); 
    dataD(2,1) = roi6.Position(1,1); dataD(2,2) = roi6.Position(1,2);
    
    plot(dataA(:, 1), dataA(:, 2), 'LineWidth',2,'Color','green');
    plot(dataB(:, 1), dataB(:, 2), 'LineWidth',2,'Color','green');
    plot(dataC(:, 1), dataC(:, 2), 'LineWidth',2,'Color','green');
    plot(dataD(:, 1), dataD(:, 2), 'LineWidth',2,'Color','green');
    
    pendienteA = (dataA(2,2) - dataA(1,2))/(dataA(2,1) - dataA(1,1)); 
    pendienteB = (dataB(2,2) - dataB(1,2))/(dataB(2,1) - dataB(1,1)); 
    pendienteC = (dataC(2,2) - dataC(1,2))/(dataC(2,1) - dataC(1,1)); 
    pendienteD = (dataD(2,2) - dataD(1,2))/(dataD(2,1) - dataD(1,1)); 
    
    lines(1).point1 = [dataA(1,1), dataA(1,2)]; lines(1).point2 = [dataA(2,1),dataA(2,2)];
    lines(2).point1 = [dataB(1,1), dataB(1,2)]; lines(2).point2 = [dataB(2,1),dataB(2,2)];
    lines(3).point1 = [dataC(1,1), dataC(1,2)]; lines(3).point2 = [dataC(2,1),dataC(2,2)];
    lines(4).point1 = [dataD(1,1), dataD(1,2)]; lines(4).point2 = [dataD(2,1),dataD(2,2)];
   
    lines(1).pendiente = pendienteA;
    lines(2).pendiente = pendienteB;
    lines(3).pendiente = pendienteC;
    lines(4).pendiente = pendienteD;
    
    dim = I1(1:256, 1:256);
    
    Name = sprintf("linesStruct_%i.mat", i);
    CutOut = I1(1:256, 1:256);
    imwrite(CutOut, Name, 'jpg'); % This is not correct. 
    
    keyboard;   
    imshow(CutOut);
    
    close all
    
end

[Walter Roberson]

Use

 Name = sprintf("linesStruct_%i.jpg", i);

if you are going to be using imwrite()

If you do not actually want to write the data as an image file, you could also save the data into a .mat file. For that you could do something like

 varname = sprintf("linesStruct_%i", i);
 clear tosave
 tosave.(varname) = CutOut;
 save('all_linesStruct.mat', '-append', '-struct', 'tosave')

This would append the data to all_lineStruct.mat as a variable named linesStruct_ followed by a number.

[David Ruiz Velasco]

Hello.

My idea consist of save each struct's data in a .mat file to be able to remember its relationship with the numbering of each image. I have got it.

I'll contextualize the situation in case I can help someone.

Previously, I developed an algorithm that does the above using the Hough transform, but to do this I had to save and load sections of the main image as the polyline grew so that the Hough transform only focuses on the region of interest. This avoids the use of post segmentation. Now I try to use that algorithm without clipping. Obviously this makes the polyline barely evolve so I try to store it in a .mat file to use designing a custom training method.

Up to this point, I can be content with saving the data independently of the image, although I understand that depending on the strategy to adopt, perhaps I should use some other storage method that allows me to associate polylines and images more directly.

Thak you very much.

[Walter Roberson]

I am not clear as to whether the problem is solved?

[David Ruiz Velasco]

Hello.

Yes, it is solved.

Thank you very much.