clc;
clear;
close all;



imds = imageDatastore('C:\DIP\sheykhi covid19\COVID-CTset',...
'IncludeSubfolders',true,'FileExtensions',{'.tif','.png'},'LabelSource','foldernames');
imds.ReadSize = numpartitions(imds)
imds.ReadFcn = @customreader;

% imds.ReadFcn = @(imds)imresize(imread(imds),[75,75,3]);



countEachLabel(imds)

for i=1:50
img = readimage(imds,i);
% imshow(img)
size(img)
% pause
end

%%plot folder
for i = 1:9
    I=imread(['000',num2str(i), '.jpg']);

    subplot(3,6,i);
    imshow(I);
end
for i = 10:18
    I=imread(['00',num2str(i), '.jpg']);

    subplot(3,6,i);
    imshow(I);
end

%%
[imdsTrain,imdsValidation] = splitEachLabel(imds,.5,'randomize');
countEachLabel(imdsTrain)

layers = [
    imageInputLayer([200 200 ])
    
    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
    
    maxPooling2dLayer(2,'Stride',2)
    convolution2dLayer(3,64,'Padding','same')
    batchNormalizationLayer
    reluLayer
    
    maxPooling2dLayer(2,'Stride',2)
    
    convolution2dLayer(3,128,'Padding','same')
    batchNormalizationLayer
    reluLayer
    
     maxPooling2dLayer(2,'Stride',2)
    
    convolution2dLayer(3,256,'Padding','same')
    batchNormalizationLayer
    reluLayer
    
    maxPooling2dLayer(2,'Stride',2)
    
    convolution2dLayer(3,512,'Padding','same')
    batchNormalizationLayer
    reluLayer
           

        
    fullyConnectedLayer(2)
    softmaxLayer
    classificationLayer];

options = trainingOptions('sgdm', ...
    'InitialLearnRate',.001, ...
    'MaxEpochs',80, ...
    'Shuffle','every-epoch', ...
    'ValidationData',imdsValidation, ...
    'ValidationFrequency',100, ...
    'Verbose',false, ...
    'Plots','training-progress');

net = trainNetwork(imdsTrain,layers,options);

%%train
YPred = classify(net,imdsTrain);
accuracy_train = sum(YPred == imdsTrain.Labels)/numel(imdsTrain.Labels)
figure
plotconfusion(imdsTrain.Labels,YPred)
[X,Y,T_train, AUC_train] = perfcurve(single(imdsTrain.Labels=='-'),single(YPred=='-'),'1');
figure
plot(X,Y, 'LineWidth',2)
xlabel('False positive rate','FontSize',14) 
ylabel('True positive rate','FontSize',14)
title('ROC for Train Data Classification','FontSize',14)


%%test
YPred = classify(net,imdsValidation);
accuracy_test = sum(YPred == imdsValidation.Labels)/numel(imdsValidation.Labels)
figure
plotconfusion(imdsValidation.Labels,YPred)
[X,Y, T_test, AUC_test] = perfcurve(single(imdsValidation.Labels=='-'),single(YPred=='-'),'1');
figure
plot(X,Y, 'LineWidth',2)
xlabel('False positive rate','FontSize',14) 
ylabel('True positive rate','FontSize',14)
title('ROC for Test Data Classification','FontSize',14)

%%All
YPred = classify(net,imds);
accuracy_all = sum(YPred == imds.Labels)/numel(imds.Labels)
figure
plotconfusion(imds.Labels,YPred)

[X,Y, T_all, AUC_all] = perfcurve(single(imds.Labels=='-'),single(YPred=='-'),'1');
figure
plot(X,Y, 'LineWidth',2)
xlabel('False positive rate','FontSize',14) 
ylabel('True positive rate','FontSize',14)
title('ROC for All Data Classification','FontSize',14)