Identifying Clump of Cells/tumors


I am trying cell profiler 2.0 to analyze few tumor images. However the images are not similar to provided in the example images and pipelines.

I have some light microscopy images of the some brain tumor cell line. I’m trying to track their growth over time.

I have attached a sample image.

I converted image to grayscale and inverted with imagemath. Then I tried to use regular treshold method to identify the object and further wanted to quantify the parameters like area, shape etc.

Can I use cell profiler for objects like these??Please let me know! Waiting for your reply!


Hi Chaitanya,

It sounds like you are doing the right initial steps. By “track”, I assume you mean simply charting the size, shape, etc, and not actually tracking the position of cells within the tumor (a somewhat more involved procedure)? If you add some measurement modules and an ExportToSpreadsheet, does that not give you output that you desire?

In general, it is much easier to quantify fluorescence images rather than brightfield, but you should be able to get gross quantifications from images like this. If not, post your pipeline, and we can better comment.


Hello Sir,

Thanks for the reply!!

You got me right on what I meant by tracking.

I’m attaching the pipeline I tried for these Images. The module “Identify Primary Objects” is not really working, may be I’m using some wrong parameters for ‘tresholding’. I just have one tumor/clump of cells in each image and all I want to do is to quantify the tumor for its area, size, shape etc.

I appreciate your help on this!

Primary_Tumors.cp (7.37 KB)


You are very close. If you change two settings in IdentifyPrimaryObjects, I think you will be fine. See the attachment for my results.
(1) Threshold correction factor 0.5 -> 1. ‘1’ is the default, and seems to work fine, as far as I can see.
(2) Lower and Upper bound, set back to 0.0 and 1.0, resp. These are also the default.

Does that look reasonable?

Thanks a ton sir!!

I tried it for one image sequence and it worked just right! I shall try few more images.

But I guess I’m good to go!