Image Processing Onramp - Course 9 - Classify images from datastore

% 1. Image Datastore
ds = imageDatastore("testimages");
nFiles = numel(ds.Files);
isReceipt = false(1, nFiles);

% 2. Classify images
for i = 1 : nFiles
    I = readimage(ds, i);
    isReceipt(i) = classifyImage(I);
end

% 3. Receipt and non-receipt images
receiptFiles = ds.Files(isReceipt)
montage(receiptFiles)
nonreceiptFiles = ds.Files(~isReceipt)
montage(nonreceiptFiles)





function isReceipt = classifyImage(I)

    gs = im2gray(I);
    gs = imadjust(gs);

    mask = fspecial("average",3);
    gsSmooth = imfilter(gs,mask,"replicate");

    SE = strel("disk",8);
    Ibg = imclose(gsSmooth, SE);
    Ibgsub =  Ibg - gsSmooth;

    Ibw = ~imbinarize(Ibgsub);

    SE = strel("rectangle",[3 25]);
    stripes = imopen(Ibw, SE);
    signal = sum(stripes,2);

    % Classification
    minIndices = islocalmin(signal,"MinProminence",70,"ProminenceWindow",25);
    Nmin = nnz(minIndices);
    isReceipt = Nmin >= 9;

end