Filling objects that share a border

Hi,

I have a question about filling in objects that share a border with the perimeter of an image.

Consider a case such as the one shown at: http://i43.tinypic.com/2nq9ua.jpg. If I run the imfill function to fill in the holes, I don’t get the five objects that touch the border of the image filled (I have pointed out the specific areas I am talking about, in this image: http://i41.tinypic.com/5afeo4.jpg).

Is there any strategy I could follow to fill holes in those areas that are outlined using the arrows?

Thanks in advance!

[Ashish Uthama]

Good question (nice image!)

http://blogs.mathworks.com/steve/2013/09/05/defining-and-filling-holes-on-the-border-of-an-image/

Answer 1

[Image Analyst]

1. Binarize and Invert the image: binaryImage = grayImage <128
2. Call imfill(binaryImage, 'holes')
3. call regionprops(binaryImage, 'BoundingBox') asking for bounding box
4. Look at the 4 sides of all the bounding boxes looking for any objects that share any bounding box side coordinate.

Give it a shot.

[Image Analyst]

OK, no response so I'm guessing that you couldn't do it, so I did it for you. Try this code (be sure to change the folder and filename):

clc;    % Clear the command window.
clearvars
close all;
workspace;  % Make sure the workspace panel is showing.
format long g;
format compact;
fontSize = 20;

% Read in a standard MATLAB gray scale demo image.
folder = 'C:\Users\Alexius\Documents';
baseFileName = '2nq9ua.jpg';
% Get the full filename, with path prepended.
fullFileName = fullfile(folder, baseFileName);
% Check if file exists.
if ~exist(fullFileName, 'file')
  % File doesn't exist -- didn't find it there.  Check the search path for it.
  fullFileName = baseFileName; % No path this time.
  if ~exist(fullFileName, 'file')
    % Still didn't find it.  Alert user.
    errorMessage = sprintf('Error: %s does not exist in the search path folders.', fullFileName);
    uiwait(warndlg(errorMessage));
    return;
  end
end
grayImage = imread(fullFileName);
% Get the dimensions of the image.  
% numberOfColorBands should be = 1.
[rows, columns, numberOfColorBands] = size(grayImage);
if numberOfColorBands > 1
  % It's not really gray scale like we expected - it's color.
  % Convert it to gray scale by taking only the green channel.
  grayImage = grayImage(:, :, 2); % Take green channel.
end
% Display the original gray scale image.
subplot(2, 2, 1);
imshow(grayImage, []);
title('Original Grayscale Image', 'FontSize', fontSize);
% Enlarge figure to full screen.
set(gcf, 'units','normalized','outerposition',[0 0 1 1]);
% Give a name to the title bar.
set(gcf,'name','Demo by ImageAnalyst','numbertitle','off') 

% Binarize and Invert the image: 
binaryImage = grayImage < 128;
% Display the image.
subplot(2, 2, 2);
imshow(binaryImage, []);
title('Binary Image', 'FontSize', fontSize);

% Call imfill(binaryImage, 'holes')
binaryImage = imfill(binaryImage, 'holes');
% Display the image.
subplot(2, 2, 3);
imshow(binaryImage, []);
title('Filled Binary Image', 'FontSize', fontSize);

% Get areas
labeledImage = bwlabel(binaryImage);
measurements = regionprops(labeledImage, 'Area')
% Sort based on areas
[sortedAreas, sortIndexes] = sort([measurements.Area], 'Descend')
% Get rid of largest one, which is the background.
% Extract only those blobs that meet our criteria, and
% eliminate those blobs that don't meet our criteria.
% Note how we use ismember() to do this.
keeperBlobsImage = ismember(labeledImage, sortIndexes(2:end));
% Display what we retained.
% Display the image.
subplot(2, 2, 4);
imshow(keeperBlobsImage, []);
title('Border blobs', 'FontSize', fontSize);

% Get bounding boxes of the border blobs.  Need to relabel again.
[labeledImage, numberOfBlobs] = bwlabel(keeperBlobsImage);
measurements = regionprops(labeledImage, 'BoundingBox', 'Centroid')
% Loop over blobs, labeling them with text.
hold on;
for b1 = 1 : numberOfBlobs
  x = measurements(b1).Centroid(1);
  y = measurements(b1).Centroid(2);
  text(x, y, num2str(b1));
end
% Loop over blobs, comparing each blob to the other blobs
for b1 = 1 : numberOfBlobs
  boundingBox1 = measurements(b1).BoundingBox;
  x1 = [boundingBox1(1), boundingBox1(1) + boundingBox1(3)]
  y1 = [boundingBox1(2), boundingBox1(2) + boundingBox1(4)]
  for b2 = 1 : numberOfBlobs
    if b2 == b1
      continue;
    end
    boundingBox2 = measurements(b2).BoundingBox;
    x2 = [boundingBox2(1), boundingBox2(1) + boundingBox2(3)]
    y2 = [boundingBox2(2), boundingBox2(2) + boundingBox2(4)]
    % See if they share the top border
    if y1(1) == y2(1)
      message = sprintf('Blob %d shares the top border with blob #%d', b1, b2);
      uiwait(msgbox(message));
    end
    % See if they share the bottom border
    if y1(2) == y2(2)
      message = sprintf('Blob %d shares the bottom border with blob #%d', b1, b2);
      uiwait(msgbox(message));
    end
    % See if they share the left border
    if x1(1) == x2(1)
      message = sprintf('Blob %d shares the left border with blob #%d', b1, b2);
      uiwait(msgbox(message));
    end
    % See if they share the right border
    if x1(2) == x2(2)
      message = sprintf('Blob %d shares the right border with blob #%d', b1, b2);
      uiwait(msgbox(message));
    end
  end
end
msgbox('Done with demo.  Thanks ImageAnalyst!');

[alexius]

thank you Image Analyst for the answer. I've tried the code, and get error..

Error in imfill (line 132)
    mask = padarray(mask, ones(1,ndims(mask)), -Inf, 'both');
Error in padarray (line 51)
binaryImage = imfill(binaryImage, 'holes');

can you help me, I’m quite new to IPT.

[alexius]

I'm sorry. I forgot to change the folder and filename. Now the code works. Thank you Image Analyst. :D

Answer 2

[Anand]

Steve Eddins just posted about this. Here's an alternative approach:

http://blogs.mathworks.com/steve/2013/09/05/defining-and-filling-holes-on-the-border-of-an-image/