Project MONAI
Project MONAI (Medical Open Network for AI) is a community-led, open-source PyTorch framework and ecosystem for deep learning in healthcare imaging. Its mission is to promote wider adoption of AI in healthcare imaging through validated, reproducible tools, standardized workflows and open collaboration across research, clinical and industry communities.
At a Glance
- Academic medical-imaging researchers
- Clinical and hospital AI teams
- Medical-device and healthcare-technology companies
- Enterprise researchers and developers
- +2 more
AI Tools by Project MONAI
(1)
MONAI
PyTorch Framework for Medical Imaging
Discussions
No discussions yet
Be the first to start a discussion about Project MONAI
Latest News
MONAI Core 1.6.1 released with security fixes, NaViT and Hyena-based components, and a modernized build process.
MONAI Core 1.6.0 released.
MONAI Core 1.5.0 released.
Radiology AI company deepc joined MONAI, announced MONAI compatibility and joined the Deploy working group.
Products & Services
The flagship PyTorch-based training and research library: medical-imaging transforms and I/O, 2D/3D/4D data handling, networks including UNETR, SwinUNETR and VISTA-3D, losses, metrics, inferers, Auto3DSeg, MONAI Bundles and multi-GPU/mixed-precision workflows.
A server-side, AI-assisted interactive annotation engine with active learning, DeepEdit/DeepGrow, multi-user workflows and integrations for 3D Slicer, OHIF, MITK, QuPath, Digital Slide Archive, CellProfiler, DSA and CVAT.
A healthcare-native deployment ecosystem for packaging trained models as MONAI Application Packages (MAPs), with an App SDK, inference engines, Workflow Manager and Informatics Gateway for DICOM/FHIR exchange and workstation, Holoscan or Kubernetes deployment.
A community hub for reproducible pretrained models distributed using the MONAI Bundle format; the initial announcement described 15 models from partner institutions.
Market Position
MONAI positions itself as a healthcare-imaging-specialized, community-governed layer on top of PyTorch rather than a general-purpose ML framework. Its differentiation is domain-specific medical-image I/O and transforms, validated 3D architectures, reproducible Bundles/Model Zoo assets, active-learning annotation and DICOM/FHIR-aware clinical deployment; adjacent alternatives and predecessor technologies include general PyTorch/TensorFlow workflows and healthcare-imaging toolkits such as NVIDIA Clara, NiftyNet, DLTK and DeepNeuro.
Leadership
Founders
NVIDIA
Technology company and original co-founder; MONAI builds on NVIDIA Clara and other healthcare-AI tooling.
King's College London
Academic co-founder through the London Medical Imaging & AI Centre for Value Based Healthcare and the School of Biomedical Engineering & Imaging Sciences.
Executive Team
Sebastien Ourselin
Chair of the MONAI Advisory Board; Head of King's College London's School of Biomedical Engineering & Imaging Sciences at launch
Professor and biomedical-engineering/imaging-sciences leader at King's College London; quoted at launch on MONAI's scientific and open-source mission.
Jorge Cardoso
MONAI co-founder/co-founding technical leader; CTO of the London AI Centre at launch
Medical-imaging AI researcher and London AI Centre technology leader; identified in the launch coverage as the CTO who described the gap MONAI addressed.
Board of Directors
Founding Story
NVIDIA and King's College London came together during MICCAI 2019 to create a standardized, user-friendly, open-source platform for deep learning in medical imaging. The initiative addressed the lack of a comprehensive, reliable framework for reproducible healthcare-imaging research and was publicly launched in April 2020 to establish shared practices for data processing, transformations, training, evaluation and benchmarking.
Business Model
Revenue Model
MONAI is a freely available Apache 2.0 open-source community project rather than a disclosed standalone commercial company. The public materials describe free software, community contributions and ecosystem adoption; no subscription, API-usage or license revenue model is stated.
Target Markets
- Academic medical-imaging researchers
- Clinical and hospital AI teams
- Medical-device and healthcare-technology companies
- Enterprise researchers and developers
- Radiology, pathology and endoscopy teams
- Institutions collaborating through federated learning
- Medical-imaging research and reproducible model development
- Radiology segmentation, classification, detection and registration on CT/MRI
- Pathology and whole-slide-image cell/tissue analysis
- Endoscopy video segmentation and surgical-tool tracking
- AI-assisted annotation and active-learning data curation
- Federated learning across healthcare institutions
- Mayo Clinic Florida / Center for Augmented Intelligence in Imaging
- Siemens Healthineers / AI-Rad Companion
- deepc / deepcOS
- AI Centre for Value Based Healthcare hospital network