I am doing CNN image classification, but i keep getting this error ; Error using trainNetwork Unexpected input size: All observations must have the same channel and spatial Di

Question

Muhaimin 2023-1-23

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1898670-i-am-doing-cnn-image-classification-but-i-keep-getting-this-error-error-using-trainnetwork-unexpe

编辑： TED MOSBY 2024-11-18，19:50

outputFolder=fullfile('seaTech');
rootFolder=fullfile(outputFolder,'seacreaturesDataset');
categories={'Crab','Shark','Starfish','Turtle','Octopus'};
imds=imageDatastore(fullfile(rootFolder,categories),'LabelSource','foldernames');
tbl=countEachLabel(imds)
figure;
perm = randperm(500,20);
for i = 1:20
    subplot(4,5,i);
    imshow(imds.Files{perm(i)});
end
img = readimage(imds,1);
size(img)
numTrainFiles = 75;
[imdsTrain,imdsValidation] = splitEachLabel(imds,numTrainFiles,'randomize');
layers = [
    imageInputLayer([100 100 3])
    
    convolution2dLayer(3,8,'Padding','same')
    batchNormalizationLayer
    reluLayer
    
    maxPooling2dLayer(1,'Stride',1)
    
    convolution2dLayer(3,16,'Padding','same')
    batchNormalizationLayer
    reluLayer
    
    maxPooling2dLayer(1,'Stride',1)
    
    convolution2dLayer(3,32,'Padding','same')
    batchNormalizationLayer
    reluLayer
    
    fullyConnectedLayer(5)
    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)

2 个评论
显示无隐藏无

KSSV 2023-1-23

What are the dimensions of your input images? They all should be of same size. If not consdier getting them into same size using imresize

Muhaimin 2023-1-23

its [100 100 3], I already convert them using imresize function;

f='Shark';

d=ls(f);

d(1,:)=[]

mkdir('NewShark')

for i=1:size(d)

I=imread(fullfile(f,d(i,:)));

I=imresize(I,[100 100]);

imshow(I)

imwrite(I,fullfile('NewShark',strcat(num2str(i),'.png')));

end

for all the file, total of 5 subfiles containing each set of data

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

TED MOSBY 2024-11-15，8:28

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1898670-i-am-doing-cnn-image-classification-but-i-keep-getting-this-error-error-using-trainnetwork-unexpe#answer_1545598

编辑：TED MOSBY 2024-11-18，19:50

在 MATLAB Online 中打开

Hi @Muhaimin,

The error means that there might be inconsistencies in the dimensions or channels of the images within your‘imageDatastore’. Even though you resized the images to[100 100], there can be differences in the number of channels (e.g., some images might be grayscale with a single channel, while others are RGB with three channels).

For that you can:

Ensure All Images Have 3 Channels:

for i = 1:size(d) 
    I = imread(fullfile(f, d(i,:))); 
    I = imresize(I, [100 100]); 
    if size(I, 3) == 1 
        I = cat(3, I, I, I);  % Convert grayscale to RGB 
    end 
    imwrite(I, fullfile('NewShark', strcat(num2str(i), '.png'))); 

Verify All Images in the Datastore:

    for i = 1:numel(imds.Files) 
        img = readimage(imds, i); 
        if size(img, 1) ~= 100 || size(img, 2) ~= 100 || size(img, 3) ~= 3 
            error('Image %d does not match the required size or channel count.', i); 
        end 
    end 
    

Update the ImageDatastore:

    imds = imageDatastore(fullfile(rootFolder, categories), 'LabelSource', 'foldernames'); 

Hope this helps!

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。

I am doing CNN image classification, but i keep getting this error ; Error using trainNetwork Unexpected input size: All observations must have the same channel and spatial Di

2 个评论
显示无隐藏无

回答（1 个）

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

另请参阅

类别

标签

Community Treasure Hunt

I am doing CNN image classification, but i keep getting this error ; Error using trainNetwork Unexpected input size: All observations must have the same channel and spatial Di

2 个评论 显示 无隐藏 无

回答（1 个）

0 个评论 显示 -2更早的评论隐藏 -2更早的评论

另请参阅

类别

标签

Community Treasure Hunt

2 个评论
显示无隐藏无

0 个评论
显示 -2更早的评论隐藏 -2更早的评论