Count objects labeled in image

Hi all,

I identified circles and rectangular by a mark circle at centroid. Now I want to count how many object have labeled and how many object not labeled?

The result is in the attached image

Really appreciate for your help

if metric < threshold
     centroid = stats(k).Centroid;
    plot(centroid(1),centroid(2),'ko');
    count = 
  end

Answer 1

[Image Analyst]

stats = regionprops(L,'Area','Centroid','Perimeter');
allAreas = [stats.Area]
allPerims = [stats.Perimeter];
allCircularities = allPerims .^2 ./ (4 * pi * allAreas)
% Find number with circularity more than 1.5
count = allCircularities > 1.5

[Image Analyst]

For what it's worth, I'm attaching my shape recognition demos.

[Tu Nguyen]

I tried your way but all value less than 1.5. Now I am trying both ways to figure out. The method of @Faraz Hedayati I am struggling how to convert the out last output image to an actual image because the spot on the image just a plot.

Answer 2

[Tala]

I would plot the centroids as filled black circles and threshold colors smaller than 10! then you only have the centorids and length(regionprops( YourImage,'centroid')) would give you the number of rectangulars.

[Tu Nguyen]

I tried your way but I cant get the answwer. Can you help me more?

clc;
close all;
clear;

I=imread('Shapes_with_holes.png');

bw1 = imbinarize(I);
imshow(bw1)
bw = imfill(bw1,'holes');    
bw = bwareaopen(bw,1);      
imshow(bw)

[B,L] = bwboundaries(bw,'noholes');
imshow(label2rgb(L,@jet,[.5 .5 .5]))
hold on
for k = 1:length(B)
  boundary = B{k};
  plot(boundary(:,2),boundary(:,1),'w','LineWidth',2)
end



stats = regionprops(L,'Area','Centroid','Perimeter');
threshold = 1;
for k = 1:length(B)

  boundary = B{k};

  delta_sq = diff(boundary).^2;    
  perimeter = stats(k).Perimeter;

  area = stats(k).Area;

  metric = 4*pi*area/perimeter^2;


  metric_string = sprintf('%2.2f',metric);

   if metric > threshold
     centroid = stats(k).Centroid;
     
    plot(centroid(1),centroid(2),'ko');
   end
end

[Tala]

please attach your image with the centroids as black cricles.

[Tu Nguyen]

Yes, sir, I did on my comment below your comment. Or if you want the result, it is here

[Tala]

In your original question you highlighted the centroids of rectangulars. In the second image you have the centoirds of circles and rectangulars, and some objects have two centroids.

back to your original question, whatever technique you used yo find the centroid of rectangulars, change your plot option to filled circles to get this (I call this image I):

Ig=rgb2gray(I);
BW=Ig<15;
imshow(BW)

info=regionprops(BW,'Centroid');
length(info)

which spits out 13

[Tu Nguyen]

How can I fill black the marked circle?

[Image Analyst]

Not sure what you want. In Faraz's last image the circles were white. If you made them black, the whole image would be black. Or do you want to fill the original image's black dots with the surrounding color, like what you'd to with regionfill()?

Did you see my Answer below?

[Tala]

instead of

plot(centroid(1),centroid(2),'ko');

use a dot as your marker with large size.

plot(centroid(1),centroid(2),'.k', 'MarkerSize',25)

[Tu Nguyen]

Hi @Faraz Hedayati, how can I process the image with the plot on it? I make the image gray but it backs to initial image without the marked circle

[Tala]

I am getting confused about you wanna do after all :).

Did you see Image Analyst's response?

If you want to save your figure as an image you can do:

saveas(gcf,'YourImage.png'); 
% or
exportgraphics(gcf, 'YourImage.png');

You can then import that as well using

theImage = imread('YourImage.png')