How to count the number of pixels in each section of an image set?

Hello:)

I have a set of 80 dicom images of lungs and I was wondering if there is an automatic way or a simple way to count the number of pixels belonging to the lung in each section.

Could anyone help me?

I've searched everywhere and I can't find a way to calculate the pixels of each image at the same time, I only got the total number of pixels in the set but thats not what i wanted..

Any kind of help is welcome.

Thank you :)

[Matt J]

Is the problem segmenting the lungs, or is it counting the number of pixels in a segmentation mask that you already have?

[Gabriela Almeida]

Its the counting the number of pixels in a segmentation mask i already have.

[Image Analyst]

@Gabriela Almeida then you must not have discovered my Image Segmentation Tutorial in my File Exchange

Image Analyst's File Exchange

[Gabriela Almeida]

@Image Analyst I tried to find something about this but i couldnt find it on your file

Answer 1

[Matt J]

A = imread('toyobjects.png');
A = A<100 | A>150; % some binary image


% get total number of pixels across all objects
pixtotal = regionprops('table',A,'Area')

pixtotal = 4×1 table
    Area 
    _____

    10871
    11922
     8797
     1644

[Gabriela Almeida]

if i do that it gives me the same pixel value 80 times, but what i want to know its the number of pixels each frame that only belongs to the lung. I binarize my 80 images and the lung is black and the rest of the image is white and i want to know the number of pixels who are black each image

[Matt J]

invert the image so that lungs are white.

pixtotal = regionprops('table',~A,'Area')

[Gabriela Almeida]

Thats what i did but it keeps looking like this @Matt J

[Matt J]

What's wrong with that?

[Gabriela Almeida]

I have to have different values of the number of pixels in each section because in each image the lung changes because it is a lung divided into 80 frames so the number of pixels belonging to the lung should vary when I scroll through all the images @Matt J

[Matt J]

Perhaps you are accidentally feeding the code the same image every time. Whatever the problem is, it's coming from parts of the code you haven't shown us. But you can see from the simplified examples we've given you that regionprops will work.

[Gabriela Almeida]

it works :) thank you so much!

Answer 2

[DGM]

Depending on what your images are like and what you need, this may be a start. You can calculate the number of true (nonzero) pixels in an image using nnz(). If your image has more than one object, then it will include them all. If you have more than one object and you want the pixel count for each object, you can do that using bwconncomp()

A = imread('toyobjects.png');
A = A<100 | A>150; % some binary image

imshow(A)

% get total number of pixels across all objects
pixtotal = nnz(A)

pixtotal = 33234

% get number of pixels in each object
CC = bwconncomp(A);
pixperobj = cellfun(@numel,CC.PixelIdxList)

pixperobj = 1×4

10871 11922 8797 1644

[Gabriela Almeida]

@DGM Thanks for the help but I think something is not right because the number of pixels is increasing with the number of images but there are images where the lung part decreases and the number of pixels increases.