Using ImageJ for stomach segmentation in CT imaging

Hello!
New to Image. Need the help of professionals. I am trying to use ImageJ for abdominal segmentation CT imaging.
I am trying to estimate the volume of the stomach by CT pictures.

Anyone have some insight/would like to help newbie?

Thank you in advance!

Hi @Ruiu,
we do quite frequently such measurements, but I can’t unfortunately comment on how to do that in ImageJ (hopefully someone can add that later, I wish to see that myself :wink: ).
The general strategy is the following:

  1. use any suitable method to segment the object of interest
  2. count the number of voxels that form the segmented object (= N)
  3. multiply this number by the physical size (in um) of voxels (= N * x * y * z)

Alternative solution is to use the the segmented model (step 1) to generate a triangular surface of the object and measure the volume occupied by that surface (many programs that can generate the surface are also capable of doing such calculation). The surfaces are used for visualization but as additional benefit can also be measured.

Ilya

1 Like

What programs do you use instead of imagej? I found neural networks on the githab for abdominal segmentation of 3D images, but due to my bad programming relationships, I could not run them.

well, we are naturally using our own stuff mib.helsinki.fi unless there are some specific tasks that can be done with other programs, for example, in ilastik (www.ilastik.org).
Ilya

Thanks for the reference.
I really care about the segmentation of the stomach, see if your programs can be applied to solve this problem. :grinning:

you are welcome! let me know if something is not clear.