Analysis of aggregates

Hi all,

I have a quite specific problem: (this is what it is about)

I seeded cells in microwells to measure the difference in aggregation. PIctures were taken through a microscope.
Now i want to analyse the size of the aggregates. Therefore i want to outline the clump of cells then to be able to anlyse them. I have to repeat this for every clump in the picture. (I have many many more pictures to go through)
I started drawing the outlines with the freehand tool, but i would want to automate this process. The wand tool is way too rough and unprecise.
I’m still a beginner, i have not much of a clue how imageJ exactly works.
If someone of you could help me further, i’d be incredibly grateful.

Thanks in advance

Hi @waterfles

and welcome to the forum!

This looks like a task for a computer driven image analysis. To get robust and good results it would help to know a little bit more about the different images.

Are the microwells over all images a the same position (pixel coordinates)?
Do you have an image with empty microwells?

If the microwells are always at the same position you could subtract an image of the empty microwells to get rid of the background. I guess this would be the simplest way and we wont have any illumination problems.

1 Like

You can use machine learning to segment your image.

With Trainable Weka Segmentation (in Fiji: Plugins > Segmentation > Trainable Weka Segmentation), I quickly tried and got the following result:


@imagejan This is so good that this application should be highlighted in the Wiki.


Wow, that looks amazing!
I’m surely going to try that tomorrow! I’ll let you know how it worked out!

Thank you so much for your quick and helpfull replies!

Kind Regards

1 Like

Thanks alot for your help!

I tried the trainable Weka segmentation, but my result isn’t as smooth and precise as yours. Could you tell me what parameters you used or give some additional tips?

Thanks in advance

In the settings, I had added the Variance and Structure features, increased the maximum sigma to 32, and checked Homogenize classes, see also the screenshot:

Hope that helps.


it definately does! Thanks a lot!

Could you explain me your reasoning why you change those parameters? So I can gain insight in the parameters?

Any tips on maken macro’s with this?

Thanks in advance!

I know it’s been a long time since this thread. But just to follow up:

In general, I presume @imagejan has carefully reviewed the TWS documentation which explains all those options, and has used it enough to have an intuition for which features are suitable in different scenarios.

The docs are pretty math-focused and less intuition-based, but you can probably get a feel for how the classification is working by trying different combinations of features with the “Train classifier” button on a small image.

I fuzzily recollect that @iarganda eventually plans/hopes to make some kind of feature browser, so that you can preview the training results with different combinations of features. In general, the combinations are exponential, but it would be doable to at least show a preview of a few commonly useful combinations on a small window of data.

See the TWS documentation on macros for instructions.