How can I calculate areas of different segmented regions which are in the same image?

Hello all as I mentioned in the title I want to calculate the areas of the segmented regions which are in the same image. Thus far, I deleted the noises on the image and segmented three objects but I do not know how to calculate the areas of these regions, and this is the code i use :

close all

% Reads the image and names it as RGB (writing file location

% w/ file location and file extension)

% and shows it

figure

subplot(2,3,1)

RGB = imread('C:\Users\hp.user\Desktop\1\image #1.png');

imshow(RGB);

% Firstly, converts the image to gray, after that converts it

% to pure black and white

I = rgb2gray(RGB);

bw = imbinarize(I);

subplot(2,3,2)

imshow(bw);

% binarize function has global threshold level, to change

% this threshold level, we can implemet :

% level = graythresh(I)

% level = 0.5

% bw = imbinarize(I, level);

% global threshold level is 0.6863

% in our image global threshold level is erasing the some

% figures on the image so the threshold level is changed

% to 0.65

level = graythresh(I);

level = 0.65;

bw = imbinarize(I,level);

subplot(2,3,3)

imshow(bw)

% Removing the noise by using morphology functions,

% remove pixels which do not belong to the objects of

% interest

% bwareaopen removes all objects contatining fewer that 36 pixels but not

% removing the noise in the objects

bw = bwareaopen(bw,36);

subplot(2,3,4)

imshow(bw);

% to remove the noise in the objects we do the following code

% fill a gap in the pen's cap

% strel('disk',R) creates a disk-shaped structuring elemnt,

% where R speccifies the radius

se = strel('disk',2);

% bw1 = imclose(bw,se) performs morphological closing on the

% grayscale or binary image bw, returning the closed

% image, bw1

bw = imclose(bw,se);

% fill any holes, so that region props can be used to estimate

% the area enclosed by each of the boundaries

% imfill(bw,locations) performs a flood-fill operation on background pixels

% the input binary image starting from the points specified in

% locations

bw = imfill(bw,'holes');

subplot(2,3,5)

imshow(bw)

% find the boundaries

% calculate boundaries of regions in image and overlay the boundaries

% on the image

[B,L] = bwboundaries(bw,'noholes');

subplot(2,3,6)

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

-------------------------------------

According to my lecturer, it is not available to use " regionprops" where image includes more than one segmented region to calculate area. Waiting for your helps. Thanks!

Answer 1

[Brian Hannan]

Hi, Fevzi. You've already done the hard work here. The cell array B contains the boundaries of your polygons. One way to calculate the area of each boundary is to count the number of pixels it contains. Some functions you might want to check out are inpolygon, poly2mask, and polyarea.

Answer 2

[Image Analyst]

Your lecturer is completely wrong. Or perhaps you didn't understand exactly what he was trying to say. You certainly CAN measure multiple blobs with one call to regionprops().

See my Image Segmentation Tutorial http://www.mathworks.com/matlabcentral/fileexchange/?term=authorid%3A31862&sort=downloads_desc

[Junaid Dar Dar]

I want to measure area of specific region in binary image, e.g., I have region1 white, region2 black, region3 white, region4 black and white. I want to measure area of region 4 ?

[Image Analyst]

Please post your image in a new question and post other ones you might have also. For example do you have other images and do they all look like that? Is the 4th blob ALWAYS the right most blob, in which case you can simply do

[labeledImage, numBlobs] = bwlabel(binaryImage); % Identify distinct blobs.
rightBlob = ismember(labeledImage, numBlobs); % Take the right-most one into a new image.
props = regionprops(rightBlob, 'all'); % Measure only the right-most blob.

[Junaid Dar Dar]

Thank you for your reply.

I have posted image in new question. All images donot look like same and 4th blob is not always the right most blob. I want to konw the area of any specific region in binary image. I wrote region4 as an example.

You can see new question in link:

Mathworks

[Wenqi Chen]

How can I measure different cells in different areas like this? And how to deal with cells on the boundaries?Thanks!

[Image Analyst]

@Wenqi Chen you can turn your boundaries into masks with poly2mask. Then use regionprops to measure the area of the region.

See my Image Segmentation Tutorial in my File Exchange:

Image Analyst's File Exchange

It's a generic, general purpose demo of how to threshold an image to find blobs, and then measure things about the blobs, and extract certain blobs based on their areas or diameters.

If you have any more questions, start your own discussion thread (rather than continue here).

Answer 3

[Fevzi Ustundag]

Thanks for the answers, I have just solved the problem after I uploaded the problem here. And you may be right about my lecturer. Maybe I did not understand him very well. Now, I am designing a GUI which contains different operations which includes also this area calculation. By the way, the areas of the objects are not seem exactly accurate. But it gives same results with MATLAB's Image Region Analyzer. Is it exact value or approximate value that acquired on Image Region Analyzer?

And this is the code I use to show areas of different objects in one figure:

figure

subplot(2,3,1)

RGB = imread('C:\Users\hp.user\Desktop\alan.png');

imshow(RGB);

I = rgb2gray(RGB);

bw = imbinarize(I);

subplot(2,3,2)

imshow(bw);

level = graythresh(I);

level = 0.65;

bw = imbinarize(I,level);

subplot(2,3,3)

imshow(bw)

bw = bwareaopen(bw,36);

subplot(2,3,4)

imshow(bw);

se = strel('disk',2);

bw = imclose(bw,se);

bw = imfill(bw,'holes');

subplot(2,3,5)

imshow(bw)

[B,L] = bwboundaries(bw,'noholes');

subplot(2,3,6)

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');

threshold = 0.94;

for k = 1:length(B)

boundary = B{k};

area = stats(k).Area;

area_string = sprintf('%2.2f',area);

if area > threshold

      centroid = stats(k).Centroid;

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

end

    text(boundary(1,2)-35,boundary(1,1)+13,area_string,'Color','k',...

      'FontSize',14,'FontWeight','bold');

end