Row-wise labelling with bwlabel() and regionprops

Hi, I am trying to label a collage row-wise with bwlabel() and regionprops and then crop indivudial images from that collage.

Currently I have tried using either extrema or centroids. Extrema has worked out significantly better for me.

For extrema my code is as follows:

for n = 1:binary_images
    binary_Images=fullfile(image_folder_binary, filenames_bin(n).name) ; % its will specify images names with full path and extension
       our_images_binary  = imread(binary_Images); % read images
       
     % Do a "hole fill" to get rid of any background pixels or "holes" inside the blobs.
       our_images_binary = imfill(our_images_binary, 'holes');
       
       L = bwlabel(our_images_binary,8)
       props = regionprops(L, our_images_binary, 'all');
       
       extrema = vertcat(props.Extrema);
       left_most_top = extrema(1:8:end, :);
       [sortedEX, sort_Order] = sortrows(fliplr(left_most_top));
       % Re-sort props 
       props = props(sort_Order);
       numberOfBlobs(n) = length(props);
       blobMeasurements{n}  = props;
       
end

and for centroid:

for n = 1:binary_images
    binary_Images=fullfile(image_folder_binary, filenames_bin(n).name) ; % its will specify images names with full path and extension
       our_images_binary  = imread(binary_Images); % read images
       
     % Do a "hole fill" to get rid of any background pixels or "holes" inside the blobs.
       our_images_binary = imfill(our_images_binary, 'holes');
       
       L = bwlabel(our_images_binary,8)
       props = regionprops(L, our_images_binary, 'all');
       
       xyCentroids = vertcat(props.Centroid);
       y = xyCentroids(:, 2); % Get all the y values (row centroids of blobs)
       [sortedY, sortOrder] = sort(y, 'ascend'); % Do the sort.
       % Re-sort props in order of increasing y.
       props = props(sortOrder);
       numberOfBlobs(n) = length(props);
       blobMeasurements{n}  = props;
       
end

I've also attached the binary image and the labelled image using the extrema, any suggestion on how to label this image row-wise?

Answer 1

[yanqi liu]

reference resources

《计算机视觉与深度学习实战》

clc; clear all; close all;
im = imread('binary image.png');
bw = im2bw(im);
sz = size(im);
bwt = imclose(bw, strel('line', round(sz(2)*0.2), 0));
[L, num] = bwlabel(bwt);
stats = regionprops(L);
rects = cat(1, stats.BoundingBox);
[~, ind] = sort(rects(:, 2));
figure; imshow(im, []);
for i = 1 : num
    bwi = bwt;
    bwi(L~=ind(i)) = 0;
    bwi = logical(bwi.*bw);
    statsi = regionprops(bwi);
    ceni = cat(1, statsi.Centroid);
    ceni = sortrows(ceni, 1);
    hold on;
    for j = 1 : size(ceni,1)
        text(ceni(j,1), ceni(j,2), sprintf('%d-%d', i, j),'Color','r')
    end
end

[no zoop]

Pretty awesome peice of code to label the binary images. However, when cropping it ends up cropping the entire row (12 rows croped) instead instead of each individual blob in a row in the entire image (~194 blobs). Do you have any suggestions? Below is the code I use to crop blobs.

message = sprintf('Would you like to crop out and save each individual images?');
reply = questdlg(message, 'Extract Individual Images?', 'Yes', 'No', 'Yes');
% Note: reply will = '' for Upper right X, 'Yes' for Yes, and 'No' for No.
if strcmpi(reply, 'Yes')
   for n = 1 : non_binary_images
       for k = 1 : number(n)      % Loop through all blobs.
           % Find the bounding box of each blob.
           thisBlobsBoundingBox  = blobMeasurements{n}(k).BoundingBox;  % Get list of pixels in current blob.
           % Extract out this blob into it's own image.
           subImage = imcrop(OG_images{n}, thisBlobsBoundingBox); 
           % Saving the cropped image
           folder = 'folder';
           thisBaseFileName = sprintf('name.tif', n, k);
           thisFullFileName = fullfile(folder, thisBaseFileName);
           imwrite(subImage, thisFullFileName,'tif');   
       end
   end
end