Re-segement previously obtained ROI

Background

I have an image of cells obtained using fluorescence microscope. I am trying to analyze the images (obtain information about area of each cell, circularity, etc.). I used Weka segmentation plugin to create a resultant image “classified image” where I could use the wand tool to outline the cells and create ROIs.
My problem is since there are cells on top of each other or very close (other images), the wand tool selects multiple cells in one ROI and I can not separate the ROI to get information about individual cells and since the cells are not round I can’t use watershed. I tried using watershed for irregular features, but it didn’t separate the cells as I wanted.
My question is: Is there a way to separate already obtained ROI? or is there a better way to get an ROI surrounding each individual cell. Keep in mind I have a lot of images to analyze so doing it by hand is out of the question

Maybe the marker controlled watershed gives you better results:https://imagej.net/Marker-controlled_Watershed

The watershed is performed with a gradient image rather than the distance and thus the shape of the object.

I get good results with that in irregular shaped cells where I have a marker for the cytoplasm.

Hello Ahmed,
In CellProfiler, you could use IdentifyPrimaryObjects (IDP) to detect the cells and trying different declumping methods. In general, separating overlapping cells could be challenging.
If you have nuclei marker, you could use those in IDP and use this cell marker in the SecondaryObject detection.

1 Like