3D images surface area computing

Image Analysis
, ,
#1

Hi everyone,
I started to work on ImageJ some months ago on 2D images from confocal microscopy so I’m a beginner on this software. Now I need to analyse 3D images but I’m lost in the many plugins and I d’ont have experience in 3D image analysis.

I made a stack of images of fat globules in an emulsion and I want to know the total volume of all globules in order to compare it with the volume of my solution.

Does somebody could explain me how to get this result ?
Is it possible to compute the total contact area between my solution and the fat globules ?

I hope you can help me.

Kind regards,
Emilien

#2

Welcome to the forum, @Emilien!

Others may have better suggestions, but I suggest checking out the 3D ImageJ Suite of @ThomasBoudier.

However, be aware that 3D analysis is substantially more complex than 2D analysis. So I’d also ask whether you are certain that you really need to do analysis in 3D. I am no expert at 3D analysis, so maybe the following ideas have problems, but:

  1. If the globules are relatively spherical and not too close together, you may be able to get away with analyzing projections?
  2. Or perhaps you could use TrackMate to identify the globules in 3D, segment them from their centers, and then extrapolate volume from max diameter?
  3. Or easier, since all you want is total volume ratio, might it be enough to simply segment the globules planewise, compute the area ratio of each plane, and then average them across all planes?

The point is that if you can avoid needing to do full-blown 3D segmentation, you will probably save yourself a lot of effort.

2 Likes
#3

Thank you for the leads you gave me, I will study them and try to apply them.

#4

Hi @Emilien,

For this particular question you do not need to care about 3D. The difficult part is to identify your globules against the background. If you can use a threshold (Image/Adjust/Threshold), you just create a binary image stack and then count the number of white and black pixels by looking at the histogram.
If your globules are more complex to segment you may need other tools.

Best,

Thomas

1 Like
#5

Thank you, I will try this !