Select folders one by one using imageDatastore function

Dear all,

I have a "Parent_Folder" and inside it I have another 5-subfolders that contain images inside them: Folder_1, Folder_2, Folder_3, Folder_4, Folder_5.

By using "imageDatastore" function, I want to make a loop in order to chose folders one by one, something like this:

for i = 1: 5

%in the 1st loop when i = 1

Group_A = Folder_1

Group_B = Folder_2, Folder_3, Folder_4, Folder_5

%in the 2nd loop when i = 2

Group_A = Folder_2

Group_B = Folder_1, Folder_3, Folder_4, Folder_5

%in the 3rd loop when i = 3

Group_A = Folder_3

Group_B = Folder_1, Folder_2, Folder_4, Folder_5

and so on.

Any idea how to make such a loop?

best regards,

Mesho

Answer 1

[Sai Pavan]

Hello,

I understand that you have a "Parent_Folder" and want to have five sets of "Group_A" and "Group_B" subfolder groups such that "Group_A" has one subfolder and "Group_B" has the rest of the four folders to store the images inside them in a "imageDatastore" with the help of their paths.

Please refer to the below code snippet that illustrates this task:

parentFolderPath = 'Parent_Folder';
subfolders = {'Folder_1', 'Folder_2', 'Folder_3', 'Folder_4', 'Folder_5'}; % List of subfolder names

for i = 1:length(subfolders)
    % Group_A will be the current subfolder
    groupAPath = fullfile(parentFolderPath, subfolders{i});
    Group_A = imageDatastore(groupAPath, 'IncludeSubfolders', true, 'LabelSource', 'foldernames');
    % Group_B will be all other subfolders
    groupBPaths = subfolders([1:i-1, i+1:end]); % Exclude current folder
    groupBPaths = fullfile(parentFolderPath, groupBPaths); % Full paths
    Group_B = imageDatastore(groupBPaths, 'IncludeSubfolders', true, 'LabelSource', 'foldernames');
end

Please refer to the below documentation to learn more about:

• "fullfile: function: https://www.mathworks.com/help/matlab/ref/fullfile.html
• "imageDatastore: function: https://www.mathworks.com/help/matlab/ref/matlab.io.datastore.imagedatastore.html

Hope it helps!

[Mesho]

It works, thank you very much.