what should i do with this error on cross validation?

i newer to matlab and i want to perform cross validation on my image dataset and i use below code for it:

path_to_images = "*\Rice_Image_Dataset";

image_datastore = imageDatastore(path_to_images, "IncludeSubfolders", true, "LabelSource", "foldernames");

% Split the data into train, validation, and test sets

[train, validation, test] = splitEachLabel(image_datastore, 0.6, 0.2, 0.2, 'randomized');

numFolds = 5;

cv = cvpartition(train.Labels, 'KFold', numFolds);

nets = cell(1, numFolds);

accuracies = zeros(1, numFolds);

for i = 1:numFolds

trainIdx = training(cv, i);

valIdx = test(cv, i);

trainData = subset(train, trainIdx);

valData = subset(train, valIdx);

rsz_train = augmentedImageDatastore([224 224 3], trainData);

rsz_val = augmentedImageDatastore([224 224 3], valData);

augmentedTrainDatastore = augmentedImageDatastore([224 224 3], trainData, 'ColorPreprocessing', 'gray2rgb');

augmentedValDatastore = augmentedImageDatastore([224 224 3], valData, 'ColorPreprocessing', 'gray2rgb');

opts = trainingOptions("sgdm", ...

"ExecutionEnvironment", "auto", ...

"InitialLearnRate", 0.01, ...

"MaxEpochs", 5, ...

"MiniBatchSize", 64, ...

"Shuffle", "every-epoch", ...

"ValidationFrequency", 70, ...

"Plots", "training-progress", ...

"ValidationData", rsz_val, ...

"Momentum", 0.9);

[net, traininfo] = trainNetwork(augmentedTrainDatastore, lgraph_1, opts);

nets{i} = net;

true_val_labels = valData.Labels;

pred_val_labels = classify(net, augmentedValDatastore);

accuracies(i) = mean(true_val_labels == pred_val_labels);

end

% Compute average accuracy over all folds

averageAccuracy = mean(accuracies);

but i recieved the following error:

Array formation and parentheses-style indexing with objects of class 'matlab.io.datastore.ImageDatastore' is not allowed. Use objects of class

'matlab.io.datastore.ImageDatastore' only as scalars or use a cell array.

could you plz help me with that.

Answer 1

[Aniketh]

Hi Amir, to resolve the error, modify the following lines in your code:

trainData = subset(train, trainIdx);
valData = subset(train, valIdx);

to

trainData = subset(train, trainIdx);
valData = subset(train, valIdx);

By using image_datastore instead of train, you ensure that the subset function operates on the imageDatastore object as intended.

Hope that helped your case!