MATLAB Answers

How to count white pixels in Watershed segmented image?

3 views (last 30 days)
I am using watershed method to this image:
And got the output:
Now I want to convert the segmented part within the yellow boundary to black and white, and calculate its white and black pixels within the boundary. How can I do it?
The part of the code is as follows:
[B,L]=bwboundaries(tumor,'noholes'); %draw boundaries around the tumor
hold on
for i=1:length(B)
plot(B{i}(:,2),B{i}(:,1), 'y' ,'linewidth',1.45);
title('Segmented Tumor');
hold off;

Accepted Answer

Image Analyst
Image Analyst on 5 Aug 2021
What is the intensity that distinguishes between your "black" and "white"? 128? Don't use image as the name of your variable -- it's already the name of a built-in function.
Anyway to get the pixel values for blob in L:
pixelValues = grayScaleImage(L == 1); % A 1-D vector of all the pixel values in the region.
% Then sum up black and white
threshold = 128; % Whatever you want.
numBlackPixels = sum(pixelValues < threshold)
numWhitePixels = sum(pixelValues >= threshold)
Image Analyst
Image Analyst on 5 Aug 2021
Try this:
thisBoundary = B{1}; % Whatever boundary you want.
x = thisBoundary(:, 2);
y = thisBoundary(:, 1);
redMask = poly2mask(x, y, rows, columns);
pixelsInRedMask = binaryImage(redMask);
numBlack = sum(pixelsInRedMask == 0);
numWhite = sum(pixelsInRedMask == 1);

Sign in to comment.

More Answers (1)

darova on 4 Aug 2021
Here is an example with using bwlabel
A = zeros(100);
t = linspace(0,2*pi,1000); % a lot of points
[x,y] = pol2cart(t,1); % circle coordinates
ind = sub2ind(size(A),round(y*15)+30,round(x*25)+50); % convert coordinates into [rows,cols]
A(ind) = 1; % fill [rows,cols] with ones
A1 = imdilate(A,ones(2)); % make line thicker (bwlabel doesn't work for thin lines)
[B,n] = bwlabel(~A1); % separate each region
imshow(B,hot(2)) % show separated region
line(x*25+50,y*15+30) % show original curve





Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!