Cell counts, diameters using particle sizer


#1

Hi all,

I’ve found this forum helpful in my attempts to quantify the number and size (diameter) of cells in my images of oyster gonad, however I’m about near the end of my rope and posting as a last ditch effort before doing a manual approach.

I’ve tried a number of different things (cellprofiler, ilastik, etc), but it seems to me the particlesizer plugin has the most promise and has gotten me closest to what I’m looking for. I’m attaching a picture before and after running particlesizer with ellipse fitting mode. As you can see, it does a decent job, but I don’t find it accurate enough yet. If only I could combine the “training” aspect of ilastik’s density counting with particlesizer’s ability to measure objects of different sizes…

Any suggestions or feedback are welcomed, and thanks to all of you who have helped others in the past and left good advice.


#2

@jlmatt

Have you looked into using Trainable Weka Segmentation(TWS) plugin? It takes advantage of machine learning techniques - similar to ilastik. See my quick attempt with the default settings below… That could get you the segmentation you need - and then use ParticleSizer to process/measure.

There is a helpful workshop for Segmentation in Fiji that goes over how to use TWS with a real world example… with slides here.

Hope this helps!

eta :slight_smile:


#3

etarena,

Thanks for your reply! Yes, I’ve found the TWS plugin and I’ve found it very useful in order to determine the overall area occupied by the particles of interest. I’ve now discovered the probability maps you can generate from TWS, which I think is getting me even closer to proper segmentation and identification of particles with particlesizer (before and after pictures below). Wondering if anyone has suggestions in how I could do even better in having particlesizer identify cells among the clumps in the middle-just using the default settings for ellipse fitting mode in particlesizer at the moment.