How to store the segmented results in a container in itk?

I’m segmenting the dicom slices, I want to store the results in one container. So that I can use them for 3d reconstruction.

Have you considered loading a series of slices as a 3D image, then create a segmentation in the form of another 3D image (usually called a label-map)? If not, what container did you have in mind?

1 Like

By container do you mean file? There are a number of file formats and variants.

What kind of images did you segment? CT/MRI/PET/US, pathology, or not medical images?

Can segments overlap?

Besides ITK, what software are you going to use for analysis, quantification, and visualization?

Do you want to use the files for long-term archival and sharing with groups outside your institution, or just for your group’s internal use for a single project?


Thanks for the response,
I segmented on 2d DICOM on 24 slices and want to store them in the memory and I need to do 3d reconstruction.

Thanks for the response,
I wanted to store results in the form of objects.
I used CT scans.

Also want to know how to perform segmentation on 3d volume?

A library forum is not the best place to start learning about image segmentation. Maybe take a look at some of the courses which employ ITK in order to get started?


I would recommend to store both your reconstructed 3D volumes and segmentations in NRRD file format. I did not get direct answers to my questions, but it seems that you are new to the field and therefore you just starting to learn about image segmentation by yourself. For this, you don’t need complicated formats, such as DICOM Segmentation Object, but you can work with the simplest research file formats, without embedding segment name&color information in the file.

The page that @dzenanz linked to above contains excellent resources, so I would recommend to check them out. Usually, it is hard to make image segmentation fully automatic - most often you need to provide some inputs and may need to review the end result and make some corrections. Therefore typically you segment images using an application with a GUI and dedicated segmentation tools, such as 3D Slicer (this is a good page to get started), MITK, ITK-Snap, etc.


Thanks for the resources, I’ll look into it.

Thank you for being so helpful!