Could someone please explain how to sort objects based on intensity? I am using the uploaded pipeline and images and wish to only classify as cells those which express a high intensity of MAP2 (file ending in c1 and used to identify secondary object)? I have tried altering the threshold strategy and levels, number of classes for thresholding and even having the program create a “corrected” image that blacks out all the cells below a given intensity. However, I have not been able to exclude the faded non-neuronal cells that should be background (such as those in the middle right of the bottom half of the image). Thank you!
PROTOTYPE_nucleus_cellsize.cpproj (105 KB)
Please see my attached, edited pipeline. The basic idea is to:
(1) Find all nuclei (you already did this)
(2) Measure the intensity of all the nuclei, using the Map2 channel
(3) DisplayDataOnImage: Show the values simply to help you set a threshold in the next step. So this step could be disabled later
(4) Filter out the objects with low mean intensity. You could certainly choose another feature (median intensity, etc), but I chose the mean.
(5) ID Secondary only on the Map2 positive nuclei.
It may not be perfect, but it’s at least a lot better than before! You will of course run into trouble if the glia (I assume they are glia) lie right next to the neuron soma and thus would have an artificially high Map2 intensity. You could add further filter criteria (size constraints, etc) to help. Or even use a machine learning approach like CellProfiler Analyst, but that is probably not worth it here.
Hope that helps!
PROTOTYPE_nucleus_cellsize-2_DLogan.cppipe (15 KB)