Different histogram output from same image

When executing this code:

locationTrain = 'D:\Kuliah\PCB\Tubes\Images\training\*.png';
train = imageDatastore(locationTrain); 
locationGTTrain = 'D:\Kuliah\PCB\Tubes\GT\training\*.png';
GTTrain = imageDatastore(locationGTTrain); 
trainName = ["017" "032" "036" "037" "040" "042" "049" "057" "060" "063" "064" "066" "068" "069" "080" "081" "084" "088" "094" "098" ];
count = 0; 
countmasked = 0; 
i = 1; 
while hasdata(train)  
    img = read(train) ;               % read image from datastore 
    red = img(:,:,1);              % Red channel 
    green = img(:,:,2);             % Green channel 
    x = read(GTTrain);
    Mask = logical(x);  % Crop Image 
    props = regionprops(Mask, 'BoundingBox'); 
    thisBB = props.BoundingBox; 
    MaskCrop = imcrop(Mask, thisBB); % Crop GT in binary 
    TrainCrop = imcrop(green, thisBB); % Crop image in red channel 
    % Masked & cropped image 
    % Use GT as mask 
    Maskedgreen = bsxfun(@times, uint8(TrainCrop), cast(MaskCrop,class(uint8(TrainCrop))));  
    figure; 
    imhist(Maskedgreen); 
    saveas(gcf,['histo green train ' num2str(trainName(i)) '.jpg']) 
    count = count + imhist(green); 
    countmasked = countmasked + imhist(Maskedgreen); 
    i = i + 1; 
end

The histogram output is:

However, I am unable to run that code and thus modifying it to:

locationTrain = 'D:\Kuliah\PCB\Tubes\Images\training\*.png';
train = imageDatastore(locationTrain); 
locationGTTrain = 'D:\Kuliah\PCB\Tubes\GT\training\*.png';
GTTrain = imageDatastore(locationGTTrain); 
trainName = ["017" "032" "036" "037" "040" "042" "049" "057" "060" "063" "064" "066" "068" "069" "080" "081" "084" "088" "094" "098" ];
count = 0; 
countmasked = 0; 
i = 1; 
while hasdata(train)  
    img = read(train) ;               % read image from datastore 
    red = img(:,:,1);              % Red channel 
    green = img(:,:,2);             % Green channel 
    x = read(GTTrain);
    Mask = logical(x);  % Crop Image 
    props = regionprops(Mask, 'BoundingBox'); 
    thisBB = props.BoundingBox;
    BB2 = reshape(thisBB, [], 2);
    thisBB = reshape(BB2(1:2,:), 1, []); 
    MaskCrop = imcrop(Mask, thisBB); % Crop GT in binary 
    TrainCrop = imcrop(green, thisBB); % Crop image in red channel 
    % Masked & cropped image 
    % Use GT as mask 
    Maskedgreen = bsxfun(@times, uint8(TrainCrop), cast(MaskCrop,class(uint8(TrainCrop))));  
    figure; 
    imhist(Maskedgreen); 
    saveas(gcf,['histo green train ' num2str(trainName(i)) '.jpg']) 
    count = count + imhist(green); 
    countmasked = countmasked + imhist(Maskedgreen); 
    i = i + 1; 
end

and the resulting histogram is:

Why can it be different? What approach can I use to produce the same histogram as before?

Answer 1

[Cris LaPierre]

Without your images to test with, it's hard to say for certain. However, the only difference in your code is the following two lines

BB2 = reshape(thisBB, [], 2);
thisBB = reshape(BB2(1:2,:), 1, []); 

This could potentially change what coordinates are used for your bounding box, which then affects what data is cropped. It would make sense, then, that if the size of the bounding box changes, the resulting histogram will change as well.

[Walter Roberson]

That section of code is from a previous response I made to the user, and had to do with compensating for the situation in which GTTrain is potentially 3 dimensions such as if it is rgb or a volume.