how to train a CNN by a folder of 2816 images?

Question

ghazaleh 2024-4-30

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2113551-how-to-train-a-cnn-by-a-folder-of-2816-images

回答： Harsh 2024-5-2

i have this code:

digitDatasetPath = fullfile(matlabroot,'toolbox','nnet','nndemos', ...

'nndatasets','DigitDataset');

digitData = imageDatastore(digitDatasetPath, ...

'IncludeSubfolders',true,'LabelSource','foldernames');

figure;

perm = randperm(50,20);

for i = 1:20

subplot(4,5,i);

imshow(digitData.Files{perm(i)});

end

CountLabel = digitData.countEachLabel;

img = readimage(digitData,1);

size(img)

trainingNumFiles = 9;

rng(1) % For reproducibility

[trainDigitData,testDigitData] = splitEachLabel(digitData, ...

trainingNumFiles,'randomize');

layers = [imageInputLayer([148 210 1])

convolution2dLayer(5,20)

reluLayer

maxPooling2dLayer(2,'Stride',2)

fullyConnectedLayer(5)

softmaxLayer

classificationLayer()];

layers = [imageInputLayer([148 210 1])

convolution2dLayer(5,20)

reluLayer

maxPooling2dLayer(2,'Stride',2)

fullyConnectedLayer(5)

softmaxLayer

classificationLayer()];

options = trainingOptions('sgdm','MaxEpochs',15, ...

'InitialLearnRate',0.0001);

convnet = trainNetwork(trainDigitData,layers,options);

YTest = classify(convnet,testDigitData);

TTest = testDigitData.Labels;

accuracy = sum(YTest == TTest)/numel(TTest)

How to train this network with a file containing 2816 png images? What changes should be made to it?

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Harsh 2024-5-2

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2113551-how-to-train-a-cnn-by-a-folder-of-2816-images#answer_1451096

在 MATLAB Online 中打开

Hi,

Based on what you've shared, it looks like you're working on training a CNN with images stored in the folder “matlabroot\toolbox\nnet\nndemos\nndatasets\DigitDataset”.

This folder contains images for 10 different classes, each image being 28 x 28 in size. To match these specifications, kindly adjust the layers variable in your code so that the image input is set to [28 28 1], and the fully connected layer is set to 10 as shown below:

layers = [imageInputLayer([28 28 1]) 
    convolution2dLayer(5,20) 
    reluLayer 
    maxPooling2dLayer(2,'Stride',2) 
    fullyConnectedLayer(10) 
    softmaxLayer 
    classificationLayer()]; 

Additionally, it seems like there is a duplicate declaration of the layers variable (line 28) of your code. Kindly remove the second instance of this declaration.

I hope this helps, thanks!