Image J Fiji 3D object counter very slow

I am using fiji image J 3D object counter to count number of cells and size within a tumor within a confocal microscope image. In the images with very large tumors or a lot of z stacks, image J is taking HOURS. I am using a mac book air to do this. Is it normal for it to take so long? or do I need a computer with higher processing power? Is there anything I can do to speed it up?


If the data is high resolution, that doesn’t sound too surprising. Usually you will want to test/fine tune parameters on a small subset of the data, then run it for the rest.

Generally segmentation will be single threaded (unless you create sub-cubes of your area), so the best you can do is find the computer with the fastest single core speed, and maybe overclock it. Also, if you don’t have an NVMe drive, any system with one will likely speed up data access speed, though I’m not sure if that is a limiting condition in your case.

Note, these are general observations from using other programs, I don’t have experience with ImageJ’s version specifically.

It might be that the Particle Analyzer of the BoneJ plugin can speed up your process because it is an optimized version of the 3D objects counter, see:

See also these topic links:

1 Like

Have you considered doing your computation using your GPU with #clij ? You’ll find all the information about that plugin here:
It works with a very wide range of GPUs and can tremendously speed up this kind of calculations. In your case, within the clij plugin, you should use:

Labelling -> Connected component analysis

If you just need the number of objects, you can then just find the maximum label in the resulting map. If you need more information on the objects (size, intensity etc.) you can then also use:

Measure -> Statistics of label map

Of course it might be that your images are too large for your GPU, so I can’t guarantee that this will work, you’ll just have to try.