Looking for imaging processing expert to assembly back cropped images

Question

Abdussalam Elhanashi 2020-11-3

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/634614-looking-for-imaging-processing-expert-to-assembly-back-cropped-images

在 MATLAB Online 中打开

Hi

I am looking for how to assebmly back cropped images

I trained the neural network and used training images and test images with cropped ones

I need in (Reconstructed images from test images) to have assembel back the cropped images as one image for each

Herein the code

close all;
clc; 
 
%%  Initalize the data 
 
dataDir= fullfile('Data/');
exts = {'.jpg','.png','.tif','BMP'};
 
 
imds = imageDatastore(fullfile(dataDir),...
'IncludeSubfolders',true,'FileExtensions','.jpg','LabelSource','foldernames');
countEachLabel(imds);
 
[TrainData, TestData] = splitEachLabel(imds, 0.5);
size(TrainData);
countEachLabel(TrainData);
numImages = numel(TrainData.Files); 
dx = 25;
dy = 25;
w = dx - 1;
h = dy - 1;
for i = 1:numImages
    img = readimage(TrainData, i);
    img=rgb2gray(img);
    img3= im2double(img); 
    scale = 0.5;
    img8 = imresize(img3,scale);
    for x = 0:3
        for y = 0:3
            ofs_x = 1 + x * dx;
            ofs_y = 1 + y * dy;
            img_crop = imcrop(img8,[ofs_x ofs_y w h]);
            %img4 = imshow(img_crop, 'InitialMagnification',800);
            %drawnow;
            idx = (i-1)*16 + y*4 + x + 1;
            % 1  2  4  5
            % 6  7  8  9 
            % 10 11 12 13
            % 14 15 16 17
            
            Train{idx} = (img_crop);         %#ok<SAGROW>
        end
    end
 
end
hiddenSize = 50;
autoenc = trainAutoencoder(Train,hiddenSize,'MaxEpochs',2000,...
'DecoderTransferFunction','purelin','EncoderTransferFunction','satlin','L2WeightRegularization',0.004,'SparsityRegularization',4,'SparsityProportion',0.15);
numImages = numel(TestData.Files); 
for i = 1:numImages
    img = readimage(TestData, i);
    img=rgb2gray(img);
    img3= im2double(img); 
    scale = 0.5;
    img8 = imresize(img3,scale);
    for x = 0:3
        for y = 0:3
            ofs_x = 1 + x * dx;
            ofs_y = 1 + y * dy;
            img_crop = imcrop(img8,[ofs_x ofs_y w h]);
            %img4 = imshow(img_crop, 'InitialMagnification',800);
            %drawnow;
            idx = (i-1)*16 + y*4 + x + 1;
            % 1  2  4  5
            % 6  7  8  9 
            % 10 11 12 13
            % 14 15 16 17
            
            Test{idx} = (img_crop);         %#ok<SAGROW>
        end
    end
 
end
xReconstructed = predict(autoenc,Test);
 %% Test Images 
figure();
for i = 1:20
    subplot(4,5,i);
    imshow(TestData.Files{i});
end
 
 %% Reconstructed images from TestData
figure();
for i = 1:20
    subplot(4,5,i);
    reconstructed = xReconstructed{i};
    reconstructed(imbinarize(reconstructed)) = 1;
    imshow(reconstructed)
 end

I would like to assemble back the cropped images as one image for each in this part from the code above

%% Reconstructed images from TestData
figure();
for i = 1:20
    subplot(4,5,i);
    reconstructed = xReconstructed{i};
    reconstructed(imbinarize(reconstructed)) = 1;
    imshow(reconstructed)
end