How to find clusters in binary 3D image?

Hello

I have a binary 3D image, i.e. it contains only 0 and 1. Now I want to find all clusters of 1s (i.e. clusters of voxels containing only value 1). Finally for each cluster I should know the coordinates of the voxels belonging to that cluster.

How can this be done? Of course I can iterate over all voxels but the difficulty is to detect the clusters and extract all voxels inside the cluster.

Answer 1

[Image Analyst]

Use bwlabeln() or bwconncomp(). If you really need the (x,y,z) coordinates of each voxel in a labeled blob, then you can call regionprops() and ask for PixelIdxList.

[Sepp]

Thank you very much for your answer.

The binary 3D image is of size 60 x 90 x 10. With the following Matlab code I'm searching for connected voxels of 1s in the image:

    conn = 26;
    CC = bwconncomp(img, conn);
    stats = regionprops(CC,'pixellist');

The really weird thing is that stats returns coordinates which are outside of the 3D image.

For example for

    voxels = stats(6).PixelList;
    voxels(23,:)

I'm getting a 3D coordinate of (61,15,4).

What is wrong here? Regarding this issue I'm also concerned that the calculation of the connected components is wrong.

Here is the 3D image (as .mat file): https://www.dropbox.com/s/pw3yzbcx4uyhh9s/img.mat?dl=0

[Image Analyst]

You've made probably the most common mistake novice users make : confusing x and y with row and column. Remember, all matrices in MATLAB are indexed by row and column, NOT by x and y. So it's M(row, column) = M(y,x), and NOT M(x,y) because x is not the row - it's the column.

PixelList returns a x,y,z vector. You have a matrix that you say is 60x90x10 which I assume is rows by columns by slices. So the max of x is 90, the max of y is 60, and the max of Z is 10. So a coordinate of (x,y,z) = (61,15,4) is well within the range of your matrix bounds of (90,60,10).