ITK-SNAP

ITK-SNAP is a free, open-source application for segmenting structures in 3D and 4D medical images. In dentistry it is used by researchers, maxillofacial teams, orthodontic labs and students to outline teeth, jaws, airways, cysts or the mandibular canal on CBCT and CT scans and to export the result as a 3D surface. If you are looking for an ITK-SNAP download for Windows, Mac or Linux, this page covers what the program does, how dental users work with it and what to watch out for.

ITK-SNAP was originally created at the University of North Carolina, and current development is led by Paul Yushkevich and colleagues at the Penn Image Computing and Science Laboratory (PICSL), University of Pennsylvania. It is released under the GNU GPL v3. The design deliberately stays focused on one job, segmentation, which makes it easier to learn than larger platforms. Version 4.4 adds AI-assisted segmentation through nnInteractive.

ITK-SNAP at a glance

What it does

• Opens DICOM series, NIfTI and many other 3D image formats, including colour, multi-channel and 3D+time images.
• Manual segmentation with paintbrush and polygon tools in three linked orthogonal views at once.
• Semi-automatic segmentation with active contours (“snake” evolution) seeded by bubbles inside the structure.
• AI-assisted segmentation in 4.4 via nnInteractive, using point, scribble or lasso prompts; the deep-learning part runs on a GPU through the ITK-SNAP DLS service, locally or on a remote machine.
• Interpolation between slices, so you can draw a few key slices and fill the rest.
• 3D cut-plane tool for fast cleanup of segmentations in the 3D view.
• Exports segmentations as surface meshes in STL, VTK, VTP, BYU or VRML format.
• Manual and automatic image registration, for example to align two scans of the same patient.

How dental teams use it

Segment a jaw or tooth from CBCT and export STL

1. Open the CBCT DICOM folder via File > Open Main Image and select the series.
2. Adjust window and level so enamel, dentine and bone are clearly separated.
3. Start the active contour tool, choose thresholding as the pre-segmentation mode and set the lower threshold for bone or tooth density.
4. Place seed bubbles inside the structure, run the evolution and stop when the contour fills it.
5. Fix leaks with the paintbrush, then use Segmentation > Export as Surface Mesh and save as STL.

Outline a cyst or lesion for volume measurement

1. Choose a label and trace the lesion with the polygon tool on several slices.
2. Use slice interpolation to fill the gaps between traced slices.
3. Check the result in the 3D view and read the label volume under Tools > Volumes and Statistics.

Use AI-assisted segmentation (version 4.4)

1. Set up the ITK-SNAP DLS Python package on a machine with a suitable GPU, or connect to a server your institution runs.
2. In ITK-SNAP, connect to the DLS service and choose the nnInteractive tool.
3. Click, scribble or lasso the structure; the model proposes a 3D segmentation that you then correct manually.

Install and first run

• Windows: run the 64-bit self-install package and start ITK-SNAP from the Start menu.
• macOS: choose the Intel or Apple Silicon (M1 and later) build that matches your Mac, drag it to Applications and approve it on first launch if macOS asks.
• Linux: unpack the 64-bit binary archive and run the itksnap launcher from the bin folder.
• Hardware: any modern 64-bit computer handles typical CBCT volumes; more RAM helps with large field-of-view scans. The AI feature needs access to a GPU, either local or remote.
• Common gotchas: if a CBCT folder contains several series, pick the right one in the DICOM dialog. If the program fails to start on older hardware, update the graphics driver; version 4.4 uses a new OpenGL-based renderer with a software fallback.

Limitations

• No dental-specific automatic models: it does not know what a tooth is unless you guide it.
• Threshold-based active contours are less reliable on CBCT than on CT because CBCT grey values are not calibrated.
• The AI mode needs extra setup (Python package and GPU access) and sends image data to the DLS service you connect to, so use a local or institutional server for patient data.
• No implant planning, guide design or report tools.
• Mesh output usually needs smoothing and repair in a mesh editor before 3D printing.
• The official site does not present it as a regulated medical device; use it for research, models and teaching, not as your diagnostic viewer.

ITK-SNAP vs alternatives

FAQ

Is ITK-SNAP free?

Yes. ITK-SNAP is free and open source under the GNU GPL v3, with no paid edition.

Where do I get the ITK-SNAP download for Windows?

A 64-bit Windows self-install package is published for each release; the official site, itksnap.org, lists current and previous versions with release notes.

Is there an ITK-SNAP download for Mac?

Yes. There are separate macOS builds for Intel Macs and for Apple Silicon (M1 and later).

Can I install ITK-SNAP on Linux or Ubuntu?

Yes. A 64-bit Linux binary is provided; unpack it and run the launcher. No iPad version exists.

ITK-SNAP vs 3D Slicer: which should I use?

ITK-SNAP is simpler and focused on segmentation. 3D Slicer is broader and has dental AI extensions, so it is the better pick for automatic tooth segmentation.

How do I use the snake tool in ITK-SNAP?

Select the active contour tool, define a region of interest, choose a pre-segmentation mode such as thresholding, place seed bubbles and run the evolution until the contour fills the structure.

Is ITK-SNAP open source and safe to install?

Yes, the source code is public on GitHub under the GPL v3. Install it from the official site or the project’s GitHub releases to be sure you get an unmodified build.

What does ITK-SNAP stand for?

ITK refers to the Insight Toolkit, the image analysis library it is built on, and SNAP stands for SNAke Automatic Partitioning, after its active contour method.

Last checked against official sources: 26 September 2026 (developer website: itksnap.org). Versions and licence terms change — confirm on the developer's site before deploying in a clinic.