Hi, the issue may occur because MATLAB's “trainNetwork” function expects the input image to be in a 4D format: (Height, Width, Channels, Batch Size). Since your trained network was designed with an image input layer (imageInputLayer([16 16 1])), it requires the input image to match this format.
Please refer to this code for implementation on dummy data
% Define Layers
layers = [
imageInputLayer([16 16 1])
convolution2dLayer(3, 8, 'Padding', 1) % Use numeric padding instead of 'same'
reluLayer
maxPooling2dLayer(2, 'Stride', 2)
fullyConnectedLayer(10)
softmaxLayer
classificationLayer];
% Define Training Options
options = trainingOptions("sgdm", ...
"MaxEpochs", 5, ...
"MiniBatchSize", 32, ...
"Verbose", false);
% Generate Random Binary Training Data (1000 samples)
numTrain = 1000;
trainImages = randi([0, 1], 16, 16, 1, numTrain); % 4D array (Height, Width, Channels, Batch)
trainLabels = categorical(randi([0, 9], numTrain, 1)); % Labels from 0 to 9
% Train Network
net = trainNetwork(trainImages, trainLabels, layers, options);
% Test with a Random 16x16 Image
testImage = randi([0, 1], 16, 16, 1); % Binary test image
testImage = reshape(testImage, [16 16 1 1]); % Ensure 4D input format
predLabel = classify(net, testImage);
disp("Predicted Label: " + string(predLabel))
