CNN error; trainNetwork(XTrain, TTrain, layers, opts); Layer 11 is expected to have a different size

Hello everyone I am trying to implement the following program taken from:

"Deep Learning for Computer Vision with MATLAB"

https://la.mathworks.com/matlabcentral/fileexchange/58030-example-files-for--deep-learning-for-computer-vision-with-matlab--webinar?s_tid=prof_contriblnk

There they use images of 32x32 but my images are of 720x480, my code is the following already adapted but I have problems in the part: "Define the architecture of CNN"

Code: -------------------------------------------------------------------------------------------------- %% inicio clc; % Clear the command window. close all; % Close all figures (except those of imtool.) clear; % Erase all existing variables. Or clearvars if you want. workspace; % Make sure the workspace panel is showing.

cd c:\[CNN]\

%% Load image Training dataset (720x480 colour images in 10 classes) imsetTrain = imageSet('VideoSalida','recursive');

%% Display Sampling of Image Data numClasses = size(imsetTrain,2); imagesPerClass = 5; imagesInMontage = cell(imagesPerClass,numClasses); for i = 1:size(imagesInMontage,2) imagesInMontage(:,i) = ... imsetTrain(i).ImageLocation(randi(imsetTrain(i).Count, 1, ... imagesPerClass)); end

montage({imagesInMontage{:}},'Size',[numClasses,imagesPerClass]); title('Ejemplo de imagenes de Entrenamiento Etiquetadas por el experto medico')

%% Prepare the data for Training % Read all images and store them in a 4D uint8 input array for training, % with its corresponding class

trainNames = {imsetTrain.Description}; XTrain = zeros(480,720,3,sum([imsetTrain.Count]),'uint8'); TTrain = categorical(discretize((1:sum([imsetTrain.Count]))',... [0,cumsum([imsetTrain.Count])],'categorical',trainNames));

j = 0; tic; for c = 1:length(imsetTrain) for i = 1:imsetTrain(c).Count XTrain(:,:,:,i+j) = read(imsetTrain(c),i); end j = j + imsetTrain(c).Count; end toc;

%% Define a CNN architecture conv1 = convolution2dLayer(5,32,'Padding',2,... 'BiasLearnRateFactor',2); conv1.Weights = gpuArray(single(randn([5 5 3 32])*0.0001)); fc1 = fullyConnectedLayer(1905,'BiasLearnRateFactor',2); fc1.Weights = gpuArray(single(randn([1905 955])*0.1)); fc2 = fullyConnectedLayer(7605,'BiasLearnRateFactor',2); fc2.Weights = gpuArray(single(randn([7605 3805])*0.1));

layers = [ ... imageInputLayer([480 720 3]); conv1; maxPooling2dLayer(3,'Stride',2); reluLayer(); convolution2dLayer(5,480,'Padding',2,'BiasLearnRateFactor',2); reluLayer(); averagePooling2dLayer(3,'Stride',2); convolution2dLayer(5,960,'Padding',2,'BiasLearnRateFactor',2); reluLayer(); averagePooling2dLayer(3,'Stride',2); fc1; reluLayer(); fc2; softmaxLayer() classificationLayer()];

% Define the training options. opts = trainingOptions('sgdm', ... 'InitialLearnRate', 0.001, ... 'LearnRateSchedule', 'piecewise', ... 'LearnRateDropFactor', 0.1, ... 'LearnRateDropPeriod', 8, ... 'L2Regularization', 0.004, ... 'MaxEpochs', 10, ... 'MiniBatchSize', 100, ... 'Verbose', true);

%% Training the CNN [net, info] = trainNetwork(XTrain, TTrain, layers, opts); -------------------------------------------------------------------------------------------------- The error that gives me is the following:

Error using trainNetwork (line 140) Layer 11 is expected to have a different size.

Error in CNN (line 85) [net, info] = trainNetwork(XTrain, TTrain, layers, opts);

Caused by: Error using nnet.internal.cnn.layer.util.inferParameters>iInferSize (line 82) Layer 11 is expected to have a different size.

Can anybody help me??

Thanks in advance Angel

[Joss Knight]

Too difficult to follow. Please format your code using the formatting options available in the edit window.

[Von Duesenberg]

A good place to start debugging your model is to use the Network analyzer from the file exchange.