How can i save data from 1*5 cell array into excel file. I have 5 images in loop and i extract glcm features now i want to save each imgaes features but only the last image data is saved.

%path D:\matlab\data\Training\glossy\*.jpg
path='D:\matlab\data\Training\test\';
list=dir([path, '*.jpg']);
for x=1:length(list)
  images{x}=imread([path, list(x).name]); 
  if length(size(images{x}))==3 %check if the image I is color
    I=rgb2gray(images{x});
  end;
  offsets0 = [0 1; -1 1; -1 0; -1 -1];
  glcm1 = graycomatrix(I,'offset',offsets0);
  stats{x} = graycoprops(glcm1,{'all'});
  writetable(struct2table(stats{x}), 'test_glcmfeatures.csv')
end

[dpb]

writetable doesn't append (as you found out) unless you use it to write Excel files and use the 'Range' argument to set a location.

Probably simplest solution is to move the write until after the loop finishes and then write the whole array at once.

Answer 1

[Adam Danz]

A snippet from your code:

    for x=1:length(list)
        ...
        writetable(struct2table(stats{x}), 'test_glcmfeatures.csv')
    end

You are overwriting test_glcmfeatures.csv on every iteration of the loop.

Instead, one solution would be to change the file name and create a new file on each iteration.

    for x=1:length(list)
        ...
        writetable(struct2table(stats{x}), sprintf('test_glcmfeatures_%d.csv', x))
    end

Alternatively, if you want to write everything to the same file, try using the ' range ' (link) property. [Addendum] Or, as @dpb pointed out, you could just collect all of the stats data within the look at write it all in one file after the loop (which would be simpler than using the 'range' property).

[Anjali Acharya]

Hey Adam Danz, Thank you for your response, I will try the way you suggest but I am not sure about the 'range'how to use that.I look forward for more help when I need. Thank you

[Adam Danz]

Sounds good. Start by clicking the link I provided to read about the 'range' property. Perhaps a better idea is to move the stats data outside of the loop as @dpb mentioned.

[Anjali Acharya]

I am a beginner in matlab so I cannot figure out how to do it, could you please help me with demo code for saving features of all images in one file.

[Adam Danz]

Do you want to write all stats data to the same file or do you want to create different files at each iteration of your for-loop?

Also, could you explain what stats{x} is? Is it a matrix of numbers? Is it always the same size at each iteration? If not, how does it change?

Even better, could you either paste stats{x} into the comments or attach a mat file with stats{x}?

[Anjali Acharya]

stats{x} is 1*5 cell array in the context as i have 5 images right now,cell array consists of glcm features of each image.

<<

>>

[Anjali Acharya]

[dpb]

How big is the image? graycoprops returns a structure of the various statistical properties, 'all' means you'll have four output fields for each image of length depth. What do you intend to do with the .csv file that could end up being sizable as .csv is not very efficient for storing data?

If it's just for saving and re-reading to do further processing, .mat files are much simpler to write/read, are far faster and smaller and retain full numerical precision, besides.

[Anjali Acharya]

For now I have features value saved in csv file for last 5th image only in this format as shown in figure below:

[Anjali Acharya]

Hello #dbp i want to use these features for training a classifier ad use it for texture classification. I will resize all image to some fix size may be 512*512 before GLCM features.