MATLAB: Trying to classify images with a CNN but getting errors

classificationimage analysisimage processingMATLABneural networkneural networks

Hi all,
Apologies in advance, I'm new to Matlab. I'm trying to pass some images to a CNN for classification but am stuck in resolving a particular error. The error is as follows: 
Error using activations
Expected layer to be one of these types:
numeric
Instead its type was nnet.cnn.layer.Layer.
Error in nnet.internal.cnn.util.validateNetworkLayerNameOrIndex (line 26)
    validateattributes(layerNameOrIndex, {'numeric'},...
Error in DAGNetwork/activationsSeries (line 263)
            layerID = nnet.internal.cnn.util.validateNetworkLayerNameOrIndex(layerID, this.Layers, 'activations');
Error in SeriesNetwork/activations (line 779)
            Y = this.UnderlyingDAGNetwork.activationsSeries(X, layerID, varargin{:});
My code is as follows: 
AnisotropyDatasetPath = fullfile(matlabroot,'Training', 'Anisotropy');
IsotropyDatasetPath = fullfile(matlabroot,'Training', 'Isotropy');
FillerDatasetPath = fullfile(matlabroot,'Training', 'Filler');
TrainingDatasetPath = fullfile(matlabroot,'Training');
cropDatasetPath = fullfile('C:\Users\ezxtg4\Downloads\JPEG pics', 'crops');
imds = imageDatastore(TrainingDatasetPath, 'IncludeSubfolders',true,...
    'LabelSource','foldernames');
labelCount = countEachLabel(imds)
numTrainFiles = 999;
[imdsTrain,imdsValidation] = splitEachLabel(imds,numTrainFiles,'randomize');
layers = [
    imageInputLayer([227 227 3])
    
    convolution2dLayer(3,8,'Padding','same')
    batchNormalizationLayer
    reluLayer
    
    maxPooling2dLayer(2,'Stride',2)
    
    convolution2dLayer(3,16,'Padding','same')
    batchNormalizationLayer
    reluLayer
    
    maxPooling2dLayer(2,'Stride',2)
    
    convolution2dLayer(3,32,'Padding','same')
    batchNormalizationLayer
    reluLayer
    
    fullyConnectedLayer(3)
    softmaxLayer
    classificationLayer];
options = trainingOptions('sgdm', ...
    'InitialLearnRate',0.01, ...
    'MaxEpochs',4, ...
    'Shuffle','every-epoch', ...
    'ValidationData',imdsValidation, ...
    'ValidationFrequency',30, ...
    'Verbose',false, ...
    'Plots','training-progress');
net = trainNetwork(imdsTrain,layers,options);
YPred = classify(net,imdsValidation);
YValidation = imdsValidation.Labels;
accuracy = sum(YPred == YValidation)/numel(YValidation)
testImage = imread('C:\Users\ezxtg4\Downloads\JPEG pics\crops\crop 1.jpeg');
testLabel = imdsValidation.Labels(1)
ds = augmentedImageDatastore([227 227 3], testImage, 'ColorPreprocessing', 'gray2rgb');
imageFeatures = activations(net, ds, layers, 'OutputAs', 'columns');
predictedLabel = predict(classifier, imageFeatures, 'ObservationsIn', 'columns')
Any ideas on how to resolve this please?

Best Answer

The variable ‘net’ already has the information regarding the layers. The function’s third input is expected to a numeric index or a character vector as has been mentioned here. For example, if you want the activation of the fourth layer, the input value should be 4. Alternatively, each layer has a name property and this property value, which is a character vector, can also be passed as an input parameter.
Hope this helps!
Related Question