medInria

medInria is a free, open-source desktop application from Inria, the French national research institute for digital science, for viewing and processing medical images, aimed at researchers, clinicians in research settings and students. It covers 2D, 3D and 4D viewing, registration, filtering, manual segmentation, meshing and diffusion MRI with tractography. A medInria download today gives you the stable 4.0.2 release from December 2025 for Windows, macOS and Linux, with a 5.0 beta available for testing.

medInria has no dental module. It is relevant to dental research because it imports DICOM, including CT and CBCT series, and offers solid general tools: volume rendering, reslicing, paint and polygon segmentation, denoising and scan-to-scan registration. Inria states plainly that it is not CE marked or FDA approved, so it is for research and teaching, not diagnosis.

medInria at a glance

What it does

• A local patient database in the Browser workspace; imports data from the file system, DICOM, PACS servers and the Shanoir research database.
• Reads any image format supported by ITK, and displays VTK meshes and line sets.
• 2D/3D views side by side or as superimposed layers, slice navigation and 3D/4D volume rendering.
• Segmentation with a paintbrush, closed polygon regions of interest and landmark-based tools, plus 3D cropping.
• Registration: block-matching linear registration, Optimus rigid registration based on mutual information, and diffeomorphic demons for non-linear alignment.
• Filters from median and Gaussian blur to NL-means denoising.
• Diffusion MRI: tensor estimation, tensor-field display and fibre tractography with bundling.
• Workspaces for meshing, reformatting (cropping and reslicing) and exporting data; since 4.0, an embedded Python 3.10 interpreter for plugins.

How dental teams use it

These are general medInria tools applied to anonymised dental data for research or teaching; none of them is a dental feature.

Review a CBCT scan in 3D

1. Import the DICOM export in the Browser workspace; it appears in the local database.
2. Open it in the viewer, switch to 3D volume rendering and adjust the window to show bone.
3. Use the Reformat workspace to crop and reslice the volume along the plane you want to study.

Outline a structure

1. Load the volume and open the segmentation tools.
2. Paint the region slice by slice, or draw closed polygons on key slices.
3. Save the mask for measurement in your study or for use in other tools.

Align two scans of the same patient

1. Load the earlier scan as the fixed image and the later one as the moving image.
2. Run a rigid registration (block matching or Optimus), then check the result with superimposed layers.
3. Apply NL-means denoising first if noise makes the alignment unstable.

Install and first run

• Windows: run the 64-bit installer and follow the wizard. Windows 10 or later is required.
• macOS: open the .dmg and drag medInria into Applications; macOS 10.15 or later is required.
• Ubuntu 24.04: install the Qt 5 prerequisites listed on the official downloads page with apt, unzip the archive and start it with ./bin/medInria.sh.
• Fedora 37: install the listed Qt 5 and Mesa packages, unzip, then run the same script.
• Test data: the project provides sample diffusion and registration datasets for a first try.
• Gotcha: the Linux archives rely on system Qt 5 libraries, so install the listed prerequisites before the first launch.
• Beta: the 5.0.1 beta is offered for testing; use 4.0.2 for work you need to reproduce.

Limitations

• Not CE marked or FDA approved; for research and teaching only.
• No dental tools: no panoramic curve, nerve tracing, implant library or tooth charting.
• Many features, such as diffusion tensor imaging and tractography, are brain-oriented and irrelevant to dentistry.
• Linux builds are packaged for Ubuntu 24.04 and Fedora 37 only; other distributions may need a source build.
• Documentation is mostly a GitHub wiki and feature pages, less detailed than commercial manuals.

medInria vs alternatives

SMILI is lighter and better at quick surface-model work, while medInria has stronger registration and filtering. BioImage Suite Web runs in a browser but needs NIfTI input. For simple CBCT viewing, Sante DICOM Viewer and Mango are quicker to learn. 3D Slicer is the broader research platform if you need extensions. Developers who want medInria-style processing in their own code use ITK, which medInria is built on.

FAQ

Where does a medInria download come from?

The installer is free to download. The official site med.inria.fr is the reference for releases, documentation and platform requirements, and previous versions are kept on the project’s GitHub.

Is medInria on GitHub?

Yes. The source code is published in the medInria/medInria-public repository, which also hosts release builds and the wiki.

Is medInria free?

Yes. medInria is free and open source under a BSD-style licence from Inria, including for commercial use under the licence terms.

Does medInria run on Mac?

Yes. There is a macOS .dmg for macOS 10.15 or later.

Can medInria open dental CBCT scans?

It imports DICOM, so standard CBCT exports can be viewed, resliced and segmented. It has no dental-specific tools and is not approved for diagnosis.

What is new in medInria 4.0?

Version 4.0 added an embedded Python 3.10 interpreter for plugins, improved stability and DICOM handling, and updated the Paint, Reslice and Polygon ROI toolboxes.

Last checked against official sources: 29 September 2026 (developer website: med.inria.fr). Versions and licence terms change — confirm on the developer's site before deploying in a clinic.