Select particles that are marked with multi-point selection

Hey,

for each picture I analyze as follows:

  1. threshold/watershed => analyze particles , which results in many circle-like particles
  2. multipoint selection on the original image (automated by “Find Maxima”)

Both analysis methods do not overlap 100%. I only want the particles which also have a point selection to analyze their area.

Each particle is added to the ROI Manager individually, the Find-Maximum-Selection is also added but all points are included in 1 ROI-Manager entry.

I thought to extract the coordinates from each Find-Maximum point in two arrays (one for x- and one for y-coordinates) and compare it to the coordinates (again x- and y-coordinates) given for each pixel of the analyzed particle. However, I have ~400 points by find maximum and almost the equal amount of analyzed particles for each pictures. If I would compare each Find-Maximum-Dot with every single pixel from every single particle, this would take too long I think.

I would need to exclude the particles that do not have a Find-Maximum-point and delete it from the ROI Manager…

Is there any quick workaround that you could think of? I know there is a function like if(matches()) and if(contains()), but I would not know how to apply this to this problem.!

Attached a graphical representation of the problem.

Thank you very much!

.imagej_forum_problem

If your regions are binary, create an empty 8bit image and set all the selection points to 255. Presumably Restore Selection and Draw would do this. Then use Binary Reconstruct (plugins in the Morphology collectionn in my page or via Fiji in the Morphology Update site). Use the original image as mask and the points image as seed. That would retain the regions in the mask that intersect with the seed points.

1 Like

Works like a charme!

Thank you very much!
What I did now:
Threshold + Watershed gives you the binary image.
Select the Find-Maxima-Points from the ROI Manager and let it display on the binary image.

Plugins => MorphoLibJ => Interactive Morphological Reconstruction (Dilation, connectivity 8)
Inversion gives you the black image back without particles that were not marked by Find-Maxima.