Using neural network in GUI to predict images

Hi all,

I'm using neural network with alexnet to predict a set of images with matlab 2017a. I've been testing the predict part with the images from the example page (https://uk.mathworks.com/help/vision/examples/image-category-classification-using-deep-learning.html) but when it comes to predict the example, I get this error when loading the training Features:

Error using SeriesNetwork>iDataDispatcher (line 525) For an image input layer, the input data for predict must be a single image, a 4D array of images, or an imageDatastore with the correct size.

This part is executed when I press a button after I load and resize an image, as shown in the code below:

% --- Executes on button press in identif.
function identif_Callback(hObject, eventdata, handles)
% hObject    handle to identif (see GCBO)
% eventdata  reserved - to be defined in a future version of MATLAB
% handles    structure with handles and user data (see GUIDATA)
global foto net featureLayer trainingSet trainingLabels
img1 = imread(foto);
img2 = imresize(img1, [227,227]);
%%%%
trainingFeatures = activations(net, trainingSet, featureLayer, ...
    'MiniBatchSize', 32, 'OutputAs', 'columns','ExecutionEnvironment','cpu'); %Here I get the error
%%%%
%%%%
classifier = fitcecoc(trainingFeatures, trainingLabels, ...
    'Learners', 'Linear', 'Coding', 'onevsall', 'ObservationsIn', 'columns');
%%%%
imageFeatures = activations(net, img, featureLayer, 'ExecutionEnvironment','cpu');
%%%%
label = predict(classifier, imageFeatures)

But using it in matlab it does work. So my question is, Do I have to create a separate function to do this?

Thanks in advance and the code of the program is attached.

[sergivaqx]

Hi all, After checking for a while, I saw the the files loaded at the beginning of the program (trainingSet.m and trainingLabels.m) are loaded as struct, so to use their contents I have to call them as trainingSet.trainingSet and trainingLabels.trainingLabels. At the moment I'm testing it...