Open-source software for automatic brain tumor segmentation and standardized clinical reporting from MRI.
Raidionics segments brain tumors (pre- and postoperative) from MRI volumes, computes tumor characteristics, and generates a standardized clinical report — all through a single graphical interface. It's built by SINTEF Medical Image Analysis and is used both as a standalone desktop app and as a 3D Slicer plugin.
Under the hood, the app is a thin front end over a set of independent, reusable backend libraries — each usable on its own (Python package, CLI, or Docker), and each documented in its own repository.
flowchart TB
subgraph FrontEnds["Front ends"]
A["Raidionics<br/>(Desktop app)"]
B["Raidionics-Slicer<br/>(3D Slicer plugin)"]
end
subgraph Backends["Backend libraries (pip / CLI / Docker)"]
C["raidionics_rads_lib<br/>pipeline orchestration + reporting"]
D["raidionics_seg_lib<br/>segmentation & classification"]
E["validation_metrics_computation<br/>cross-validation & metrics"]
end
F["Raidionics-models<br/>trained model zoo and test resources"]
A --> C
B --> C
C --> D
C --> F
D --> F
E --> F
| Repository | Role | Use it directly if you want to... |
|---|---|---|
| Raidionics | Desktop application (GUI) | Just use the software, no coding |
| Raidionics-Slicer | 3D Slicer plugin | Work inside 3D Slicer |
| raidionics_rads_lib | Orchestrates full segmentation + reporting pipelines | Run an end-to-end pipeline (segmentation → report) from Python/CLI/Docker |
| raidionics_seg_lib | Segmentation/classification inference (ONNX Runtime) | Run just the segmentation step, or integrate it into your own pipeline |
| validation_metrics_computation | K-fold cross-validation and segmentation metrics | Evaluate your own models against ground truth |
| Raidionics-models | Pretrained model collection and test data | Browse or download models used by the above |
| AeroPath | Airway segmentation benchmark dataset | Benchmark airway segmentation methods |
| LyNoS | Multilabel lymph node segmentation dataset (contrast CT) | Benchmark lymph node segmentation methods |
Note on repo ownership: the three backend libraries currently live under the maintainer's personal account (
dbouget) rather than theraidionicsorg, while the front ends, models, and datasets are underraidionics. Both are part of the same project.
- Just want to use the software? → Download Raidionics for Windows, Ubuntu (≥18.04), macOS, or macOS ARM.
- Working in 3D Slicer? → Install the Raidionics-Slicer plugin.
- Building your own pipeline or research code? → Start with
raidionics_seg_libfor segmentation alone, orraidionics_rads_libfor full pipelines including reporting. - Validating a model? →
validation_metrics_computation.
Each backend library can be installed via pip, run as a CLI, called as a Python module, or run in Docker — see the respective repo's README for details.
If you use any part of Raidionics in your research, please cite:
Main software release (pre- and postoperative segmentation, standardized reporting):
Bouget, D., Alsinan, D., Gaitan, V., Holden Helland, R., Pedersen, A., Solheim, O., & Reinertsen, I. (2023). Raidionics: an open software for pre- and postoperative central nervous system tumor segmentation and standardized reporting. Scientific Reports, 13. doi:10.1038/s41598-023-42048-7
Preliminary validation (preoperative segmentation methodology):
Bouget, D., Pedersen, A., Jakola, A.S., et al. (2022). Preoperative Brain Tumor Imaging: Models and Software for Segmentation and Standardized Reporting. Frontiers in Neurology, 13. doi:10.3389/fneur.2022.932219
Full BibTeX and CITATION.cff metadata are available in each individual repository.
Most repositories are distributed under the BSD-2-Clause license (datasets AeroPath/LyNoS use MIT) — see each repository for its specific license.