calculating centroids of multiple ROIs at a time

Hi,

I have segmented my ROIs, which are circles/ellipses. I need to calculate the centroids of all ROIs at a time and save those values in Excel file for further processing. I am not able to figure out how to do that.

I am attaching my segmented image with multiple ROIs.

Thanks in advance.

Answer 1

[DGM]

If you have the binarized image, just use regionprops.

inpict = imread('blobs.png');  % standard test image

S = regionprops(inpict,'centroid')

S = 28×1 struct array with fields:
    Centroid

vertcat(S.Centroid)

ans = 28×2

7.2857 7.2857 5.0102 130.9389 28.0000 19.5263 48.2878 111.2950 38.0556 44.5556 42.5345 246.2414 51.7201 180.0177 52.6038 31.5094 35.5000 69.5000 36.5000 65.5000

Note that I'm not using your image in this example. I don't want to deal with cropping all the padding off, and the size (and thus the coordinates) are probably altered now anyway. Don't save images by saving the displayed figure. There's rarely a good reason to degrade images by doing that. Just use imwrite().

[Sunetra Banerjee]

Hi DGM,

Thank a lot for your help.

[Sunetra Banerjee]

Hi DGM,

Using regionprops function I am getting the centroids in ascending order. How to get them sequentially?

[DGM]

Sequential with respect to what? Regionprops() should be ordering them as bwlabel does. See:

https://www.mathworks.com/matlabcentral/answers/342893-how-regionprops-orders-the-regions

If you want them sorted by some metric of particular use in your application, you can use regionprops to get extra information to facilitate sorting.

For example, if I wanted to sort them by distance from the origin or by area:

inpict = imread('blobs.png');  % standard test image
S = regionprops(inpict,'centroid','area');

% sort by distance
c = vertcat(S.Centroid);
d = sqrt(c(:,1).^2 + c(:,2).^2); % get distance
[d idx] = sort(d,'ascend'); % sort distance
c = c(idx,:) % sort centroids based on prior sort

c = 28×2

7.2857 7.2857 28.0000 19.5263 38.0556 44.5556 52.6038 31.5094 37.0000 56.0000 45.0000 56.0000 41.5000 61.5000 36.5000 65.5000 35.5000 69.5000 42.5000 66.0000

% sort by area instead
a = vertcat(S.Area); % get area
[a idx] = sort(a,'ascend'); % sort area
c = c(idx,:) % sort centroids based on prior sort

c = 28×2

42.5000 69.5000 177.5000 39.0000 35.5000 69.5000 42.5000 66.0000 48.2878 111.2950 5.0102 130.9389 123.6500 108.1750 38.0556 44.5556 7.2857 7.2857 322.2857 265.7714

If there is some higher-level concept which dictates how the objects are ordered based on the structure of what the image represents, you'll need to find a way to describe it in terms of object properties.

[Sunetra Banerjee]

Hi,

For my work I need the centroids in ascending order of Y axis. If you have any advice, I would really appreciate it.

Thanks

[Sunetra Banerjee]

Thanks a lot. I am able to sort it as i need. I followed this link: https://www.mathworks.com/matlabcentral/answers/342893-how-regionprops-orders-the-regions