How to detect objects with intensity 1 in background 0

So I’m not finding a way to identify objects that are simply the binary mask of other objects from which the outlines have been subtracted.

I’ve found quite sourprising that I can’t find objects that are completly white in a completly black background. All objects are separated by the completly black boundaries. In same places there are few objects and in others quite a lot ot them. That may be turning the smoothing a litle bit crazy but I set it to 0 or 1 and had no luck. I tried every possible thresholding but didn’t help much.


Hi Felipe,

Have you tried using the manual method in IdentifyPrimaryObjects and setting the threshold value to an absolute number? If you use 0.5, that should work. The one caveat is that there may be some smoothing behind the scenes in IdentifyPrimaryObjects; if so, and the objects aren’t being detected as they should, then use a lower value.

You should also turn off discarding objects based on size or touching the image edge, as well as setting the declumping method to “None,” and unchecking the “Fill holes” option. This will serve the preserve the objects without attempting to apply the usual segmentation techniques to the binary image.

If that doesn’t work for you, upload an image and a pipeline, and we’ll take a look.

At the end I did it increasing the distance between the objects (subtracting thicker borders) for identify primary and then find the good limits with identify secondary :smile: