How to normalize overall area, and anyway to select large area to manually de-object

cellprofiler

#1

Hi,

I identify nucs, use them as a seed for the secondary object. so far so good. The goal is to count the number of secondary objects (using very stringent (1.6) conditions).

I have 2 issues:

  1. A large swathe of each overall image (brain slices) is not wanted. However, there is no way to crop it out during acquisition. I tied manual Crop in CP, but it only gives me geometric shapes. I lose too much of the wanted area. Also, using coordinates is impossible because each slice is a bit diff.
    So I tried manual select object after secondary object, but there are hundreds of cells, to click on - not realistic to work this way,
  2. Unfortunately, during acquisition of the 30 slides, the student (I did not catch this in time), tiled sometimes 4 sometimes up to 8 tiles. So counting is now a problem. Any ideas how to normalize? (Maybe measure overall image size and do some calculation from there?)

(Defined as new user on this newish forum, though I been here since 2010, so can’t upload my pipeline)

Cheers,
Avi


#2

Hi Avi_Jacob,
Just reply to this , just say hi! or anything. Then you will be able to upload anything.
Bob


#3

Hi Avi,

At least for your first issue, you could first use the module IdentifyObjectsManually to freehand draw a shape then you can freehand draw any shape then use it to Crop your image by setting the set to crop to the object you just drew.

Laura


#4

Cool thanks Laura, I’ll try that.
(I solved the normalization issue by adding a count of the nucs and ratioing)


#5

hi! or anything. (need to hit 20 char…) :slight_smile: