MITK (Medical Imaging Interaction Toolkit)

MITK (Medical Imaging Interaction Toolkit) is a free, open-source platform for viewing, segmenting, measuring and registering medical images, developed by the German Cancer Research Center (DKFZ) in Heidelberg. Its desktop application, the MITK Workbench, is used by imaging researchers, clinical research groups and students who need more than a simple DICOM viewer. If you are looking for an MITK download, the current release is MITK 2026.06, published in July 2026 for Windows, macOS and Ubuntu.

MITK has no dental edition and no dental modules. What it does offer a dental researcher is solid general tooling: it opens CT and CBCT DICOM series as 3D volumes, lets you segment bone, teeth or lesions manually or with AI-assisted tools, measures distances and angles on slices, and turns a segmentation into a surface model that can be saved as STL for 3D printing. It does not plan implants or design surgical guides.

MITK at a glance

What it does

• DICOM Browser – scans folders (including subfolders) for DICOM files, keeps a local list of patients, studies and series, and can query and retrieve images from a PACS server.
• 2D, 3D and 4D display – axial, sagittal and coronal views plus a 3D window; release 2026.06 replaced volume rendering of segmentations with a faster, cleaner polygon-mesh display.
• Manual segmentation – add/subtract, lasso, paint and wipe, region growing, fill, erase, close and live wire tools, with 2D and 3D interpolation between slices and multilabel segmentations.
• Semi-automatic 3D tools – threshold, upper/lower threshold, Otsu, GrowCut and a picking tool for isolating connected “islands”.
• AI-assisted segmentation – nnInteractive (v2.5 in this release), which segments a structure from a few clicks or scribbles, and TotalSegmentator (v2.14), which predicts masks for a large number of anatomical structures in CT and MR images.
• Surface models – “Create polygon model” and “Create smoothed polygon model” run marching cubes on a segmentation; the documentation names 3D printing as a use, and surfaces can be saved as STL.
• Measurements – lines, paths, angles, four-point angles, circles and ellipses on any slice, stored together with the image.
• Registration and scripting – rigid and deformable registration through the MatchPoint views, plus a new mitk-python package and a REST API for controlling the Workbench remotely.

How dental teams use it

Open and review a CBCT scan

1. Open the DICOM Browser and use Scan directory on the folder or disc copy that holds the CBCT export.
2. Select the series and click View (or double-click it); the volume appears in the Data Manager.
3. Scroll through axial, coronal and sagittal slices, adjust the grey-value window and use the 3D view for orientation.

Segment the mandible or a lesion and export an STL

1. In the Segmentation view, create a new segmentation for the CBCT image and add a label (for example “mandible”).
2. Start with the 3D threshold tool for bone, or use nnInteractive and mark the structure with a point or scribble; clean up with the paint, wipe and close tools.
3. Right-click the segmentation in the Data Manager and choose Create smoothed polygon model.
4. Select the new surface and save it as STL for printing or for further editing in a mesh tool such as Meshmixer.

Measure on slices

1. Open the Measurement view and select the image as the reference.
2. Pick the line, path or angle tool and click the points on the slice; results are stored as child nodes of the image.
3. Save the project from the File menu to keep the image, segmentations and measurements together.

Install and first run

• Windows: MITK is offered as a 64-bit installer and as a zip archive. Windows 10 or later is required.
• macOS: there are two separate disk images, one for Apple Silicon (arm64) and one for Intel (x86_64); pick the one that matches your Mac. macOS 13 or later is required.
• Linux: a tar.gz archive built for Ubuntu 22.04 and newer.
• First launch: the Workbench opens with perspectives such as Research, Viewer and Visualization. Add views (Segmentation, Measurement, DICOM Browser) through Window → Show View.
• AI tools: nnInteractive and TotalSegmentator are installed from inside the Segmentation view on first use. They download Python packages and model weights, need several gigabytes of disk space and run best on a GPU – the documentation recommends 6 GB or more of video memory for nnInteractive and a CUDA-capable GPU for TotalSegmentator. Both can run on a CPU, only much more slowly; nnInteractive can also use a remote GPU server.
• TotalSegmentator licences: some TotalSegmentator tasks are marked “License required” and need a free non-commercial licence key before they run.

Limitations

• No dental functions: no implant libraries, no nerve-canal or panoramic workflow, no tooth charting, no surgical guide design.
• The MITK website does not describe the Workbench as a certified medical device; treat it as research software, not a diagnostic system.
• The interface is built for imaging specialists. Expect to spend time with the user manual before segmentation feels quick.
• AI segmentation is only practical with a capable GPU; on an ordinary office PC it is slow.
• Releases regularly change menus and APIs, so tutorials written for older versions may not match the current interface.

MITK vs alternatives

FAQ

What is MITK?

MITK stands for Medical Imaging Interaction Toolkit. It is both a C++ framework for building imaging software and a ready-to-use desktop application, the MITK Workbench.

Is MITK free?

Yes. MITK is free and open source under a BSD-style licence and can be used in research and commercial settings. Some optional AI models used inside it carry their own licence terms.

Is there an MITK download for Windows?

Yes. MITK 2026.06 runs on 64-bit Windows 10 and later and comes as an installer or a zip archive. Documentation and future updates are published on mitk.org.

Is there MITK for Mac?

Yes. MITK 2026.06 supports macOS 13 and later, with separate builds for Apple Silicon and Intel Macs.

MITK vs 3D Slicer: which should I use?

Both are free research platforms that open DICOM, segment structures and export surfaces. Choose 3D Slicer if you need its dental and craniofacial extensions such as SlicerCMF; choose MITK if its segmentation tools and workflow suit you better.

Can MITK plan dental implants?

No. MITK has no implant planning or guide design. It can view CBCT, segment structures, measure and export STL models.

Last checked against official sources: 28 September 2026 (developer website: mitk.org). Versions and licence terms change — confirm on the developer's site before deploying in a clinic.