--- title: "MONAI - Medical Open Network for AI" description: "MONAI is the leading open-source framework for healthcare imaging AI, trusted by researchers and clinicians worldwide. Build, train, and deploy medical AI solutions with industry-standard tools." canonical: https://project-monai.github.io/ audience: [researcher, engineer, clinician, newcomer] last_updated: 2026-09-08 source: index.html --- # Medical Open Network for Artificial Intelligence The PyTorch-based framework for medical-imaging AI: research transforms, pre-trained models, and reproducible clinical deployment in one ecosystem. Open source, community-led. [Get Started](core.html#quick-start) [View on GitHub](https://github.com/Project-MONAI) [ MONAI Label Active-learning annotation ](label.html)[ MONAI Core Training & research ](core.html)[ MONAI Deploy Clinical inference pipelines ](deploy.html) 9.5M+ pip installs 5K+ publications citing MONAI 40+ models in the Zoo 20+ challenge wins Maintained by researchers and engineers at NVIDIA, NIH, King's College London, Mayo Clinic, MSKCC, Stanford, DKFZ, and 30+ other institutions. Why MONAI ## Built for medical imaging Domain-specific transforms, validated 3D architectures, and reproducible workflows. This is not a general ML framework retrofitted for healthcare. 1. ### PyTorch native Built on PyTorch, so there is no new framework to learn. Standard `nn.Module`, `DataLoader`, AMP, and DDP all work as you expect. 2. ### Domain-specific tooling DICOM and NIfTI I/O, spatial transforms for 3D medical volumes, segmentation metrics (Dice, Hausdorff, Surface Distance), and losses calibrated for class-imbalanced anatomy. 3. ### Reproducible Bundles MONAI Bundles package weights, training configs, metadata, and inference code together. The same format powers the Model Zoo, so your work is portable from day one. 4. ### Community-governed Maintained by NVIDIA, NIH, King's College London, Mayo Clinic, MSKCC, Stanford, DKFZ, and 30+ other institutions. Governance, working groups, and roadmap decisions are public. 5. ### State-of-the-art architectures Reference implementations of MAISI (synthetic CT generation), UNETR and SwinUNETR (3D transformer backbones), VISTA-3D (universal segmentation), and Auto3DSeg (automated pipelines). 6. ### Apache 2.0 Open source under Apache 2.0. Freely usable in commercial products, research, and clinical pipelines, with no hidden licensing gotchas for production use. Ecosystem ## Three projects, one pipeline Annotate in **Label**, train in **Core**, deploy with **Deploy**. They share one bundle format and the same data conventions, so you can use any piece on its own or all three together. Annotate ### MONAI Label Server-side annotation with active learning. It suggests the next volume to label and refines predictions from your corrections. Plugs into 3D Slicer, OHIF, and MITK. - Active learning for efficient data selection - Multiple viewer integrations - AI-assisted annotation - Multi-user collaboration [Explore Label](label.html) Train ### MONAI Core The training and research library. Medical-imaging transforms, 3D architectures, losses, and metrics: the building blocks researchers reach for first. - Medical-specific transforms - MAISI, UNETR & VISTA-3D architectures - Pre-trained model zoo - Automated ML pipelines [Explore Core](core.html) Deploy ### MONAI Deploy The path from trained model to clinical pipeline. DICOM and FHIR I/O, containerized MAP packaging, and inference runtimes from workstation to Kubernetes. - Clinical workflow integration - DICOM & FHIR support - Containerized deployment - Inference optimization [Explore Deploy](deploy.html) Connect ## Community & Support Questions, contributions, and discussions are public. Ask on GitHub or Slack; attend working-group calls; present at the MONAI Bootcamp. ### Discussion Forums [ #### GitHub Discussions Technical discussions & support ](https://github.com/Project-MONAI/MONAI/discussions)[ #### Slack Channel Real-time chat & collaboration ](https://join.slack.com/t/projectmonai/shared_invite/zt-3hucgm02q-i8Bn9XofDZs2UGOH4jUl4w) ### Events & Learning [ #### YouTube Channel Tutorials, demos & presentations ](https://www.youtube.com/@projectmonai)[ #### Tutorials Repository Notebooks & learning materials ](https://github.com/Project-MONAI/tutorials) ### Get Involved [ #### Contribution Guide Learn how to contribute ](https://github.com/Project-MONAI/MONAI/blob/master/CONTRIBUTING.md)[ #### Issue Tracker Report bugs & request features ](https://github.com/Project-MONAI/MONAI/issues) Impact ## Success Stories How hospitals, vendors, and research groups run MONAI in production. ![Mayo Clinic](/assets/img/logos/mayo-clinic.png) Featured ### Clinical AI Integration at Mayo Clinic The Center for Augmented Intelligence at Mayo Clinic Florida runs MONAI-packaged models inside its clinical Radiology pipeline. The case study documents the integration pattern, MAP containerization, and throughput results. Clinical Integration Radiology AI Workflow [Read Case Study](mayo-case-study.html) [![MONAI models running in the CAII clinical viewer at Mayo Clinic: MRI-unsafe device detection on a chest X-ray, breast-density classification on mammography, white-matter disease segmentation on brain MRI, and coronary-artery stenosis detection on CTA](/assets/img/figures/mayo-case-study-figure-3-thumb.jpg) Mayo Clinic Florida · CAII viewer](mayo-case-study.html) ![Mercure](/assets/img/logos/mercure.png) ### Mercure DICOM Orchestration MONAI Application Packages run inside the Mercure DICOM Orchestrator, so a trained model can join a radiology routing workflow without custom glue code. DICOM MAP Orchestration [Read Blog](https://monai.medium.com/rapid-deployment-of-monai-application-packages-maps-in-radiology-workflows-using-the-mercure-fe7cfd77acce) ![Siemens Healthineers](/assets/img/logos/siemens.png) ### Enterprise AI at Scale Siemens Healthineers adopted MONAI Deploy for their Digital Marketplace, enabling enterprise-scale AI deployment globally. Enterprise Global Marketplace [Read Blog](https://blogs.nvidia.com/blog/rsna-siemens-healthineers-monai-medical-imaging-ai/) ![Mayo Clinic](/assets/img/logos/mayo-clinic.png) × ![Siemens](/assets/img/logos/siemens.png) ### Global AI Marketplace Mayo Clinic AI apps built with MONAI accessible to 10,000+ institutions via Siemens Digital Marketplace. 10K+ Institutions Zero-Code [Read Case Study](mayo-siemens-case-study.html) [View All Success Stories](successstories.html) Contributors ## Maintained by the medical-imaging community These institutions have dedicated engineering, clinical, and research staff to MONAI. Contributions are reviewed and merged in the open. [Join Project MONAI](https://github.com/Project-MONAI/MONAI/blob/master/CONTRIBUTING.md) ![Answer Digital](/assets/img/logos/answer-digital.png) ![CAS](/assets/img/logos/cas.png) ![DKFZ](/assets/img/logos/dkfz.png) ![FNLCR](/assets/img/logos/fnlcr.png) ![Guy's and St Thomas'](/assets/img/logos/guys-and-st-thomas.png) ![King's College London](/assets/img/logos/kcl.png) ![Kitware](/assets/img/logos/kitware.png) ![Mayo Clinic](/assets/img/logos/mayo-clinic.png) ![MGH & BWH](/assets/img/logos/mgh-bwh.png) ![MSKCC](/assets/img/logos/mskcc.png) ![NIH NCI](/assets/img/logos/nih-nci.png) ![NVIDIA](/assets/img/logos/nvidia.png) ![Stanford](/assets/img/logos/stanford.png) ![TUM](/assets/img/logos/tum.png) ![UCL](/assets/img/logos/ucl.png) ![Vanderbilt](/assets/img/logos/vanderbilt.png) ![Warwick](/assets/img/logos/warwick.png) --- --- title: "MONAI Core" description: "MONAI Core is the flagship library of Project MONAI for medical AI development. It includes medical-specific image transforms, state-of-the-art models like UNETR for 3D segmentation, and Auto3DSeg for automated model selection." canonical: https://project-monai.github.io/core.html audience: [researcher, engineer] last_updated: 2026-09-08 source: core.html --- MONAI Core · Train # The training library built for medical imaging DICOM and NIfTI transforms, validated 3D architectures (UNETR, SwinUNETR, VISTA-3D, MAISI), and losses calibrated for class-imbalanced anatomy, all on plain PyTorch, so your `nn.Module` and `DataLoader` habits carry over. [Get Started](#quick-start) [Documentation](https://monai.readthedocs.io/en/stable/) [GitHub](https://github.com/Project-MONAI/MONAI) train a 3D segmenter $ pip install monai ``` from monai.networks.nets import UNETR from monai.losses import DiceCELoss from monai.transforms import LoadImaged model = UNETR(in_channels=1, out_channels=14, img_size=(96, 96, 96)) loss = DiceCELoss(to_onehot_y=True) ``` Capabilities ## Everything between raw scans and a trained model Not a general ML framework retrofitted for healthcare: every layer, from data loading to evaluation, is designed around 3D medical volumes. ### Medical-Specific Transforms `LoadImaged`, `Spacingd`, `RandCropByPosNegLabeld`: dictionary transforms that keep image and label in sync `CacheDataset` and smart caching for up to 10x faster data loading Deterministic pipelines for reproducible training runs DICOM, NIfTI, and PNG/JPEG I/O built in (2D, 3D, and 4D) ### State-of-the-Art Architectures MAISI: latent-diffusion generation of synthetic 3D CT with controllable anatomy VISTA-3D: universal 3D segmentation, promptable across 100+ anatomical classes UNETR and SwinUNETR transformer backbones, plus classic UNet variants 40+ pre-trained Model Zoo bundles ready for fine-tuning ### Research Workflows `sliding_window_inference` for whole-volume prediction on limited GPU memory Dice, Hausdorff, and Surface Distance metrics with batched GPU evaluation Multi-GPU (DDP) and mixed-precision training out of the box TensorBoard visualization and experiment tracking hooks Quick Start ## One install, ten composable modules Import only what you need (transforms, networks, losses) or use the end-to-end workflows. Every module links to its API reference. preprocess a CT volume $ pip install monai ``` from monai.transforms import ( Compose, LoadImage, EnsureChannelFirst, ScaleIntensity, ) transforms = Compose([ LoadImage(image_only=True), EnsureChannelFirst(), ScaleIntensity(), ]) image = transforms("ct_chest.nii.gz") ``` Want the full walkthrough? [Browse the tutorials repository](https://github.com/Project-MONAI/tutorials) ### Core Modules [ #### Transforms Spatial, intensity, and dictionary-based medical transforms ](https://monai.readthedocs.io/en/stable/transforms.html)[ #### Networks UNETR, SwinUNETR, SegResNet, and more reference nets ](https://monai.readthedocs.io/en/stable/networks.html)[ #### Losses DiceLoss, DiceCELoss, FocalLoss for imbalanced anatomy ](https://monai.readthedocs.io/en/stable/losses.html)[ #### Metrics Dice, Hausdorff, Surface Distance, and more ](https://monai.readthedocs.io/en/stable/metrics.html)[ #### Data CacheDataset, PersistentDataset, efficient loaders ](https://monai.readthedocs.io/en/stable/data.html)[ #### Inferers Sliding-window and patch-based inference ](https://monai.readthedocs.io/en/stable/inferers.html)[ #### Optimizers Optimizers and LR schedules for medical tasks ](https://monai.readthedocs.io/en/stable/optimizers.html)[ #### Applications End-to-end apps: Auto3DSeg, MAISI, VISTA ](https://monai.readthedocs.io/en/stable/apps.html)[ #### Visualization 2D/3D plotting and TensorBoard utilities ](https://monai.readthedocs.io/en/stable/visualize.html)[ #### Utils Determinism, type conversion, helper functions ](https://monai.readthedocs.io/en/stable/utils.html) Automation ## Auto3DSeg: segmentation that tunes itself Point Auto3DSeg at a labeled dataset and it analyzes intensity, size, and spacing, configures candidate algorithms, trains them on your GPUs, and ensembles the winners. The same pipeline is behind multiple MICCAI challenge wins. ### How it works #### Dataset Analysis Automatic profiling of intensity ranges, volume sizes, and voxel spacing to pick the right preprocessing. #### Algorithm Generation Generates ready-to-train algorithm folders (SegResNet, SwinUNETR, DiNTS) configured from the data assessment. #### GPU-Accelerated Training Multi-GPU training, validation, and inference without extra configuration. #### Model Ensemble Combines the best-performing candidates into a single, more accurate ensemble. ### Proven in competition #### MICCAI 2023 Challenges - Multiple 1st place wins in BraTS 2023 (brain tumors) - 1st place in KiTS 2023 (kidney segmentation) - 1st place in SEG.A. 2023 (aorta segmentation) - 1st place in MVSEG 2023 (mitral valve) [Explore Auto3DSeg tutorials](https://github.com/Project-MONAI/tutorials/tree/main/auto3dseg) Impact ## The default research stack for medical-imaging AI The numbers are verifiable: each links to its source where one exists. 5K+ ### Peer-reviewed papers Published research that builds on MONAI, across radiology, pathology, and beyond. 20+ ### Challenge wins MONAI-powered entries have topped medical-imaging competitions including BraTS and KiTS. 40+ ### Model Zoo bundles Pre-trained, reproducible bundles: weights, configs, and inference code packaged together. Reference ## How to cite MONAI Core If MONAI supports your research, please cite the framework paper: @article{cardoso2022monai, title={MONAI: An open-source framework for deep learning in healthcare}, author={M Jorge Cardoso, Wenqi Li, Richard Brown, Nic Ma, Eric Kerfoot, Yiheng Wang, Benjamin Murrey, Andriy Myronenko, Can Zhao, Dong Yang, Vishwesh Nath, Yufan He, Ziyue Xu, Ali Hatamizadeh, Andriy Myronenko, Wentao Zhu, Yun Liu, Mingxin Zheng, Yucheng Tang, Isaac Yang, Michael Zephyr, Behrooz Hashemian, Sachidanand Alle, Mohammad Zalbagi Darestani, Charlie Budd, Marc Modat, Tom Vercauteren, Guotai Wang, Yiwen Li, Yipeng Hu, Yunguan Fu, Benjamin Gorman, Hans Johnson, Brad Genereaux, Barbaros S Erdal, Vikash Gupta, Andres Diaz-Pinto, Andre Dourson, Lena Maier-Hein, Paul F Jaeger, Michael Baumgartner, Jayashree Kalpathy-Cramer, Mona Flores, Justin Kirby, Lee A D Cooper, Holger R Roth, Daguang Xu, David Bericat, Ralf Floca, S Kevin Zhou, Haris Shuaib, Keyvan Farahani, Klaus H Maier-Hein, Stephen Aylward, Prerna Dogra, Sebastien Ourselin, Andrew Feng}, journal={arXiv:2211.02701}, year={2022} } [View on DOI](https://doi.org/10.48550/arXiv.2211.02701) Connect ## Build with the community Questions, contributions, and roadmap discussions all happen in the open: on GitHub, Slack, and the working-group calls. ### Documentation API reference and concept guides, from your first transform to distributed training. [Read the Docs](https://monai.readthedocs.io/en/latest/index.html) ### GitHub Repository Source code, issues, and pull requests, with contributions reviewed and merged in the open. [View Code](https://github.com/Project-MONAI/MONAI) ### Slack Community Real-time help and collaboration with maintainers, researchers, and clinicians. [Join Slack](https://join.slack.com/t/projectmonai/shared_invite/zt-3hucgm02q-i8Bn9XofDZs2UGOH4jUl4w) ### Tutorials Hands-on notebooks covering segmentation, classification, registration, and generative models. [View Tutorials](https://github.com/Project-MONAI/tutorials) [Or join the Developers Working Group](wg_developers.html) --- --- title: "MONAI Label" description: "MONAI Label is an intelligent image labeling and learning tool that uses AI assistance to reduce the time and effort of annotating new datasets. By learning from user interactions, it refines its predictions as you annotate." canonical: https://project-monai.github.io/label.html audience: [engineer] last_updated: 2026-09-08 source: label.html --- MONAI Label · Annotate # Interactive annotation for medical imaging A server-side annotation engine with active learning: it suggests the next volume to label, refines its predictions from your corrections, and plugs straight into 3D Slicer, OHIF, and MITK. DICOM and NIfTI, in and out. [Get Started](#quick-start) [Documentation](https://monai.readthedocs.io/projects/label/en/latest/) [GitHub](https://github.com/Project-MONAI/MONAILabel) start an annotation server $ pip install monailabel ``` $ monailabel apps --download \ --name radiology --output apps $ monailabel start_server \ --app apps/radiology \ --studies datasets/imagesTr \ --conf models deepedit ``` Capabilities ## AI assistance inside the viewers you already use MONAI Label runs as a server behind your viewer. Annotators keep their tools; the model keeps learning from every correction. ### AI-Assisted Annotation Interactive models (`DeepEdit` and DeepGrow) turn a few clicks into full 3D segmentations Predictions refine in real time as you correct them Fine-tunes UNet, UNETR, and SwinUNETR backbones in the background ### Clinical Integration Native plugins for 3D Slicer, OHIF, and MITK, plus QuPath, DSA, and CVAT Reads studies directly from DICOMweb / PACS or a local folder Multi-user annotation against one shared server ### Extensible Platform Apps are plain Python: swap in your own models and inference logic Pluggable active-learning strategies (random, epistemic, custom) REST API for integrating with existing labeling pipelines Use Cases ## Ready-made apps for three imaging domains Each sample app ships with pre-configured models and viewer integrations. Start annotating in your domain without writing code. ![Radiology Use Case](/assets/img/figures/label-use-case-radiology.png) ### Radiology Organ segmentation, tumor delineation, and anatomical measurement on CT and MRI, through 3D Slicer, OHIF, or MITK, with DeepEdit and Segmentation models pre-wired. [Radiology sample app](https://github.com/Project-MONAI/MONAILabel/tree/main/sample-apps/radiology) ![Pathology Use Case](/assets/img/figures/label-use-case-pathology.png) ### Pathology Cell detection and tissue classification on whole-slide images, with QuPath, Digital Slide Archive, and CellProfiler integrations for microscopy-scale data. [Pathology sample app](https://github.com/Project-MONAI/MONAILabel/tree/main/sample-apps/pathology) ![Endoscopy Use Case](/assets/img/figures/label-use-case-endoscopy.png) ### Endoscopy Polyp detection and surgical-tool tracking in video, with CVAT integration for efficient frame-by-frame annotation and automated propagation. [Endoscopy sample app](https://github.com/Project-MONAI/MONAILabel/tree/main/sample-apps/endoscopy) Quick Start ## From zero to an annotation server in four commands Install the package, grab the radiology app and a sample dataset, and start the server. Then connect from 3D Slicer or OHIF and begin annotating with AI assistance. 1 ### Install MONAI Label One pip install gets you the server, CLI, and sample-app tooling. $ pip install monailabel 2 ### Download the radiology app Pre-configured with DeepEdit and segmentation models for CT/MRI. ``` $ monailabel apps --download \ --name radiology \ --output apps ``` 3 ### Download a sample dataset The Medical Segmentation Decathlon spleen task is a good first dataset. ``` $ monailabel datasets --download \ --name Task09_Spleen \ --output datasets ``` 4 ### Launch the server Point your viewer at the server URL and start annotating. ``` $ monailabel start_server \ --app apps/radiology \ --studies datasets/Task09_Spleen/imagesTr \ --conf models deepedit ``` [Full quickstart guide, including viewer setup](https://monai.readthedocs.io/projects/label/en/latest/quickstart.html) Active Learning ## Every correction makes the next label cheaper The server scores unlabeled volumes by model uncertainty and serves the most informative one next, so annotation effort goes where it teaches the model most. ![Active Learning Framework Diagram](/assets/img/figures/label-active-learning-framework.png) 50–80% Less annotation time 2x Faster model convergence 90% Accuracy with fewer labels ### Smart Sample Selection - Uncertainty sampling surfaces the hardest cases first - Diversity metrics keep training data varied - Ensemble disagreement flags ambiguous anatomy ### Continuous Model Improvement - Models retrain on new labels while you keep annotating - Transfer learning from pre-trained MONAI bundles - UNet, UNETR, and SwinUNETR supported out of the box ### Quality Assurance - Live validation metrics during annotation sessions - Per-prediction uncertainty estimates - Automated label quality checks Learning ## Resources & Training Videos, guides, and example apps to go from first install to a custom annotation workflow. ### Video Tutorials [ #### MONAI Label Deep Dive Series Comprehensive tutorials covering all aspects of MONAI Label ](https://www.youtube.com/playlist?list=PLtoSVSQ2XzyD4lc-lAacFBzOdv5Ou-9IA)[ #### MONAI Bootcamp Overview and hands-on training from our latest bootcamp ](https://www.youtube.com/watch?v=-HAryYAO5J4) ### Documentation & Guides [ #### Quickstart Guide Get up and running with MONAI Label in minutes ](https://monai.readthedocs.io/projects/label/en/latest/quickstart.html)[ #### API Documentation Detailed technical documentation and API references ](https://monai.readthedocs.io/projects/label/en/latest/index.html)[ #### Sample Applications Radiology, pathology, and endoscopy apps with best practices ](https://github.com/Project-MONAI/MONAILabel/tree/main/sample-apps) Reference ## How to cite MONAI Label If MONAI Label supports your research, please cite the Medical Image Analysis paper: @article{monailabel2024, title={MONAI Label: A framework for AI-assisted Interactive Labeling of 3D Medical Images}, author={Diaz-Pinto, Andres and Alle, Sachidanand and Nath, Vishwesh and Tang, Yucheng and Ihsani, Alvin and Asad, Muhammad and P{\\'e}rez-Garc{\\'i}a, Fernando and Mehta, Pritesh and Li, Wenqi and Flores, Mona and Roth, Holger R. and Vercauteren, Tom and Xu, Daguang and Dogra, Prerna and Ourselin, Sebastien and Feng, Andrew and Cardoso, M. Jorge}, journal={Medical Image Analysis}, year={2024}, doi={10.1016/j.media.2024.103207} } [View on DOI](https://doi.org/10.1016/j.media.2024.103207) Connect ## Shape AI-assisted annotation Get help, share your annotation workflows, and join the Human-AI Interaction Working Group: the community standardizing how humans and models annotate together. ### Documentation From basic concepts to custom apps and advanced annotation strategies. [Read the Docs](https://monai.readthedocs.io/projects/label/en/latest/index.html) ### Human-AI Interaction WG The working group defining standard interfaces for AI-assisted annotation cycles. [Join the Group](wg_human_ai_interaction.html) ### GitHub Repository Source code, sample apps, and discussions for real-world annotation scenarios. [View Code](https://github.com/Project-MONAI/MONAILabel) ### Slack Community Real-time help and collaboration with the MONAI Label community. [Join Slack](https://join.slack.com/t/projectmonai/shared_invite/zt-3hucgm02q-i8Bn9XofDZs2UGOH4jUl4w) --- --- title: "MONAI Deploy" description: "MONAI Deploy is a framework for developing, packaging, testing, and deploying medical AI applications in clinical and research environments. Its modular architecture covers the App SDK, Workflow Manager, and Informatics Gateway." canonical: https://project-monai.github.io/deploy.html audience: [engineer] last_updated: 2026-09-08 source: deploy.html --- MONAI Deploy · Run # Your model, packaged for the clinic Wrap a trained model as a MONAI Application Package (MAP): a container that consumes DICOM, runs inference, and emits DICOM results. One artifact runs on a workstation, Holoscan, or Kubernetes. [Get Started](#quick-start) [Documentation](https://monai.readthedocs.io/projects/monai-deploy-app-sdk/en/stable/) [GitHub](https://github.com/Project-MONAI/monai-deploy-app-sdk) package & run a MAP $ pip install monai-deploy-app-sdk ``` $ monai-deploy package my_app \ -c app.yaml -t my_app:latest \ --platform x64-workstation $ monai-deploy run my_app:latest \ -i input/ -o output/ ``` Why MONAI Deploy ## Deployment built for healthcare, not retrofitted to it Generic MLOps stacks stop at the model server. MONAI Deploy speaks DICOM and FHIR, packages reproducibly, and scales from a research workstation to a hospital fleet. ### Standardized Packaging One MAP container format for every target environment Reproducible multi-site deployments: same artifact, same behavior Dependencies pinned and bundled at package time ### Healthcare-Native I/O DICOM in, DICOM/FHIR out, with PACS integration via the Informatics Gateway Built-in operators for series selection, segmentation writing, and reports Designed for clinical-workflow and compliance requirements ### Scalable Architecture Horizontal scaling on Kubernetes or single-node workstations Workflow Manager orchestrates pipelines and monitors execution Load balancing and failover for production reliability Platform ## From trained model to PACS MONAI artifacts MONAI Deploy subsystems Third-party systems ### Researchers & developers build 1 #### Trained Model From MONAI Core, MONAI Label, or any PyTorch checkpoint you can wrap in an operator. 2 #### App SDK A Pythonic SDK to compose, test, and package inference apps from reusable operators. 3 #### MAP Container Model, pre/post-processing, and dependencies in one portable, deployment-ready image. MAP handoff ### Hospital operations run 4 #### Inference Engine Executes MAPs efficiently: local runner, Holoscan, or cluster runtimes. 5 #### Workflow Manager Routes studies to the right MAP and orchestrates multi-step pipelines. 6 #### Informatics Gateway Secure DICOM and FHIR exchange with hospital information systems. 7 #### PACS Results land back in the archive radiologists already use (third-party system). ### The three subsystems, in depth ### App SDK Build & package Compose apps from operators like `DICOMSeriesToVolumeOperator` and `MonaiSegInferenceOperator`, then package them into MAPs with the `monai-deploy` CLI. [SDK Documentation](https://monai.readthedocs.io/projects/monai-deploy-app-sdk/en/stable/index.html) ### Workflow Manager Orchestrate The central orchestration service: routes incoming studies, triggers the right MAPs, tracks task state, and coordinates clinical pipelines end to end. [View Repository](https://github.com/Project-MONAI/monai-deploy-workflow-manager) ### Informatics Gateway Connect Standards-based DICOM and FHIR I/O between AI applications and hospital systems: the bridge to PACS, RIS, and EHR. [View Repository](https://github.com/Project-MONAI/monai-deploy-informatics-gateway) Quick Start ## Install, run, package, deploy Four steps take the bundled simple-imaging example from source to a runnable MAP container on your workstation. 1 ### Set up the environment Install the App SDK and clone the repository for the example apps. ``` $ pip install monai-deploy-app-sdk $ git clone https://github.com/Project-MONAI/monai-deploy-app-sdk.git $ cd monai-deploy-app-sdk $ pip install matplotlib Pillow scikit-image ``` [Setup guide](https://monai.readthedocs.io/projects/monai-deploy-app-sdk/en/stable/getting_started/installing_app_sdk.html) 2 ### Run the app locally Execute the simple imaging example straight from Python, no container yet. ``` $ python examples/apps/simple_imaging_app/app.py \ -i examples/apps/simple_imaging_app/brain_mr_input.jpg \ -o output ``` [App tutorial](https://monai.readthedocs.io/projects/monai-deploy-app-sdk/en/stable/getting_started/tutorials/simple_app.html) 3 ### Package it as a MAP Build the Docker image that runs identically in any environment. ``` $ monai-deploy package examples/apps/simple_imaging_app \ -c simple_imaging_app/app.yaml \ -t simple_app:latest \ --platform x64-workstation ``` [Packaging guide](https://monai.readthedocs.io/projects/monai-deploy-app-sdk/en/stable/developing_with_sdk/packaging_app.html) 4 ### Run the packaged container Same inputs, same outputs, now from the deployable artifact. ``` $ mkdir -p input && cp examples/apps/simple_imaging_app/brain_mr_input.jpg input/ $ monai-deploy run simple_app-x64-workstation-dgpu-linux-amd64:latest \ -i input -o output ``` [Deployment guide](https://monai.readthedocs.io/projects/monai-deploy-app-sdk/en/stable/developing_with_sdk/executing_packaged_app_locally.html) [ #### Example Applications Segmentation, classification, and DICOM-pipeline examples to copy from. ](https://github.com/Project-MONAI/monai-deploy-app-sdk/tree/main/examples)[ #### API Reference Operators, application classes, and packaging modules in detail. ](https://monai.readthedocs.io/projects/monai-deploy-app-sdk/en/latest/modules/index.html)[ #### Community Support GitHub Discussions for deployment questions and architecture advice. ](https://github.com/Project-MONAI/monai-deploy/discussions) Case Studies ## Running in production today The same MAP format scales from a single research PACS to a marketplace reaching 10,000+ institutions. ![Siemens Healthineers](/assets/img/logos/siemens.png) ### Siemens Healthineers Enterprise AI Integration MONAI Deploy is integrated into the AI-Rad Companion platform, cutting model deployment time across Siemens' global healthcare network and improving performance monitoring. Enterprise Scale AI Platform Clinical Integration [Read Blog](https://blogs.nvidia.com/blog/rsna-siemens-healthineers-monai-medical-imaging-ai/) ![Mayo Clinic](/assets/img/logos/mayo-clinic.png) ### Mayo Clinic Clinical AI Integration The Center for Augmented Intelligence in Imaging at Mayo Clinic Florida runs MONAI-packaged models inside its clinical Radiology pipeline. The case study documents the pattern. Clinical Research AI Infrastructure Imaging [Read Case Study](mayo-case-study.html) ![AI Centre for Value Based Healthcare](/assets/img/logos/aicentre.jpg) ### AI Centre for Value Based Healthcare Research Platform AIDE, the AI Deployment Engine, safely deploys AI models into research workflows across the UK's AI Centre hospital network. Research Deployment Engine AI Safety [Learn More](https://www.aicentre.co.uk/our-platforms#tab-1) ![Mercure](/assets/img/logos/mercure.png) ### Mercure DICOM Orchestration Platform An open-source DICOM orchestrator with native MAP support: drop a MONAI Application Package into a radiology research workflow without custom glue code. Open Source DICOM MAP Integration [Read Blog](https://monai.medium.com/fe7cfd77acce?source=friends_link&sk=894711683f40c61b1116fc5097a24b0a) ![Mayo Clinic](/assets/img/logos/mayo-clinic.png) × ![Siemens Healthineers](/assets/img/logos/siemens.png) ### Mayo Clinic & Siemens Digital Marketplace Scalable Research Deployment Mayo Clinic AI applications built with MONAI are available to over 10,000 institutions through the Siemens Digital Marketplace: zero-code, globally scaled deployment of MONAI-powered solutions. Global Scale Zero-Code Research [Read Case Study](mayo-siemens-case-study.html) Reference ## How to cite MONAI Deploy If MONAI Deploy supports your research, please cite the JMIR AI paper: @article{gupta2024monai, title={Current State of Community-Driven Radiological AI Deployment in Medical Imaging}, author={Gupta, Vikash and Erdal, Barbaros and Ramirez, Carolina and Floca, Ralf and Genereaux, Bradley and Bryson, Sidney and Bridge, Christopher and Kleesiek, Jens and Nensa, Felix and Braren, Rickmer and Younis, Khaled and Penzkofer, Tobias and Bucher, Andreas Michael and Qin, Ming Melvin and Bae, Gigon and Lee, Hyeonhoon and Cardoso, M Jorge and Ourselin, Sebastien and Kerfoot, Eric and Choudhury, Rahul and White, Richard D and Cook, Tessa and Bericat, David and Lungren, Matthew and Haukioja, Risto and Shuaib, Haris}, journal={JMIR AI}, volume={3}, pages={e55833}, year={2024}, doi={10.2196/55833} } [View on DOI](https://doi.org/10.2196/55833) Connect ## Deploy with the community The Deploy Working Group meets in the open. Bring your integration questions, hospital constraints, and deployment war stories. ### Documentation Concepts, tutorials, and deployment strategies for the full Deploy stack. [Read the Docs](https://monai.readthedocs.io/projects/monai-deploy-app-sdk/en/latest/index.html) ### Deploy Working Group The group steering the MAP spec, subsystems, and clinical-integration patterns. [Join the Group](wg_deploy.html) ### GitHub Repository Source for the App SDK, Workflow Manager, and Informatics Gateway. [View Code](https://github.com/Project-MONAI/monai-deploy) ### Slack Community Real-time help from maintainers and other teams deploying to the clinic. [Join Slack](https://join.slack.com/t/projectmonai/shared_invite/zt-3hucgm02q-i8Bn9XofDZs2UGOH4jUl4w) --- --- title: "MONAI Model Zoo - Pre-trained Models for Medical Imaging" description: "Explore MONAI Model Zoo - a collection of pre-trained models for medical imaging tasks. Find and use state-of-the-art models for your healthcare AI applications." canonical: https://project-monai.github.io/model-zoo.html audience: [engineer] last_updated: 2026-09-08 source: model-zoo.html --- Model Zoo # Pre-trained models for medical AI Every model ships as a MONAI Bundle: weights, training config, and inference code in one reproducible unit. The interactive browser needs JavaScript; the full catalog is listed below. - ## Brain MRI Latent Diffusion Synthesis (v1.0.3) A latent diffusion model that generates 160x224x160 voxel T1-weighted brain MRI volumes with 1mm isotropic resolution. The model accepts conditional inputs for age, gender, ventricular volume, and brain volume, enabling controlled generation of brain images with specific demographic and morphological characteristics. brain\_image\_synthesis\_latent\_diffusion\_model - ## BraTS MRI Axial Slices Latent Diffusion Generation (v1.1.4) Latent diffusion model that synthesizes 2D brain MRI axial slices (240x240 pixels) from Gaussian noise, trained on the BraTS dataset. The model processes 1-channel latent space features (64x64) and generates FLAIR sequences with 1mm in-plane resolution, capturing diverse tumor and brain tissue appearances. brats\_mri\_axial\_slices\_generative\_diffusion - ## BraTS MRI Latent Diffusion Generation (v1.1.4) Volumetric latent diffusion model that generates 3D brain MRI volumes (112x128x80 voxels) with tumor features from Gaussian noise, trained on the BraTS multimodal MRI dataset. brats\_mri\_generative\_diffusion - ## BraTS MRI segmentation (v0.5.4) 3D segmentation model for delineating brain tumor subregions from multimodal MRI scans (T1, T1c, T2, FLAIR). The model processes 4-channel input volumes with 1mm isotropic resolution and outputs 3-channel segmentation masks for tumor core (TC), whole tumor (WT), and enhancing tumor (ET). brats\_mri\_segmentation - ## Breast density classification (v0.1.8) A deep learning model for automated classification of breast tissue density in mammograms according to the BI-RADS density categories (A through D). The model processes 299x299 pixel images and classifies breast tissue into four categories: fatty, scattered fibroglandular, heterogeneously dense, and extremely dense. breast\_density\_classification - ## Chest X-ray Latent Diffusion Synthesis (v1.0.2) A latent diffusion model that generates 512x512 pixel chest X-ray images from a 64x64x77 dimensional latent space. The model processes text-based condition inputs through a 1024-dimensional context vector, enabling controlled generation of X-rays with specific pathological features. cxr\_image\_synthesis\_latent\_diffusion\_model - ## CT-CHAT (v1.1.0) CT-CHAT is a multimodal AI assistant specifically designed for 3D chest CT imaging interpretation and analysis. The model excels at tasks including visual question answering, report generation, and multiple-choice questions, leveraging full 3D spatial information for superior performance compared to 2D-based approaches. hf\_ct\_chat - ## Endoscopic In-Body Classification (v0.5.1) A binary classification model based on SENet that distinguishes between inside-body and outside-body frames in endoscopic videos. The model processes 256x256 pixel RGB images and filters irrelevant frames, enabling automated procedure analysis. endoscopic\_inbody\_classification - ## Endoscopic Tool Segmentation (v0.6.2) A 2D segmentation model that identifies and delineates surgical instruments in endoscopic video frames. The model processes 736x480 pixel RGB images and provides binary segmentation masks. Based on an EfficientNet-UNet architecture, the model supports real-time analysis of surgical procedures. endoscopic\_tool\_segmentation - ## EXAONE Path 2.0 (v1.0.0) EXAONE Path 2.0, a pathology foundation model that learns patch-level representations under direct slide-level supervision. Using only 37k WSIs for training, EXAONE Path 2.0 achieves state-of-the-art average performance across 10 biomarker prediction tasks, demonstrating remarkable data efficiency. hf\_exaonepath\_2.0 - ## EXAONEPath (v1.1.0) EXAONEPath is a patch-level pathology foundation model that achieves state-of-the-art performance across multiple pathology tasks while maintaining computational efficiency. It excels in tissue classification, tumor detection, and microsatellite instability assessment. hf\_exaonepath - ## EXAONEPath-CRC-MSI-Predictor (v1.0.0) MSI classification of CRC tumors using EXAONEPath - a patch-level foundation model for pathology. hf\_exaonepath-crc-msi-predictor - ## HoVer-Net: Nuclear Segmentation and Classification (v0.2.8) A multi-task learning model based on the HoVer-Net architecture that simultaneously performs nuclei segmentation and type classification in H&E-stained histology images. The model processes 256x256 pixel RGB patches and outputs three complementary predictions: binary nuclear segmentation (Dice score: 0.83), hover maps for instance separation, and pixel-level nuclear type classification. pathology\_nuclei\_segmentation\_classification - ## Llama3-VILA-M3-13B (v1.1.0) VILA-M3 is a medical visual language model built on Llama 3 and VILA architecture. This 13B parameter model performs medical image analysis including segmentation, classification, visual question answering, and report generation across multiple imaging modalities. hf\_llama3\_vila\_m3\_13b - ## Llama3-VILA-M3-3B (v1.1.0) VILA-M3 is a medical visual language model built on Llama 3 and VILA architecture. This 3B parameter model performs medical image analysis including segmentation, classification, visual question answering, and report generation across multiple imaging modalities. hf\_llama3\_vila\_m3\_3b - ## Llama3-VILA-M3-8B (v1.1.0) VILA-M3 is a medical visual language model built on Llama 3 and VILA architecture. This 8B parameter model performs medical image analysis including segmentation, classification, visual question answering, and report generation across multiple imaging modalities. hf\_llama3\_vila\_m3\_8b - ## Lung Nodule CT Detection (v0.6.10) A 3D detection model for identifying pulmonary nodules in CT scans. The model processes variable-sized patches and outputs detection boxes with classification scores. Trained on the LUNA16 challenge dataset, it provides automated screening capabilities for pulmonary nodule detection in chest CT examinations. lung\_nodule\_ct\_detection - ## MAISI: Medical AI for Synthetic Imaging (v1.0.2) MAISI is a diffusion-based model for generating synthetic 3D CT images with anatomical control. The model produces realistic CT volumes up to 512×512×768 voxels and can generate images conditioned on organ segmentations of 127 anatomical structures. maisi\_ct\_generative - ## Medical Image Classification Template (v0.0.4) A comprehensive template for developing 2D medical image classification models, featuring a modular architecture and standardized training pipeline. The template supports single-channel 128x128 pixel input images and outputs 4-class probability distributions, serving as a foundation for custom medical image classification tasks. classification\_template - ## Medical Image Segmentation Template (v0.0.4) A comprehensive 3D segmentation framework designed as a foundation for developing custom medical volumetric segmentation models. The template includes a configurable architecture and preprocessing pipeline, processing 128x128x128 voxel volumes with single-channel input and producing 4-class segmentation outputs. Includes support for random sphere generation for demonstration and testing purposes. segmentation\_template - ## MedNIST DDPM Hand X-ray Generation (v1.0.3) A denoising diffusion probabilistic model (DDPM) that synthesizes hand X-ray images based on the MedNIST dataset. The model learns the underlying distribution of the dataset through an iterative denoising process, demonstrating the capabilities of diffusion models in medical image synthesis. Features progressive noise-to-image generation with fine-grained control over the generation process. mednist\_ddpm - ## MedNIST GAN (v0.4.4) A generative adversarial network (GAN) that synthesizes hand X-ray images based on the MedNIST dataset. The model generates 64x64 pixel hand radiographs with varying appearances and orientations. The generated images maintain anatomical plausibility and can be used for data augmentation and educational purposes. mednist\_gan - ## MedNIST Hand X-ray Registration (v0.0.7) A ResNet-based spatial transformer model for precise registration of hand X-ray images from the MedNIST dataset. The model processes 64x64 pixel input pairs (moving and fixed images) and outputs registered images, demonstrating the application of deep learning in medical image registration. mednist\_reg - ## Multi-organ Abdominal Segmentation (v0.0.6) A 3D segmentation model optimized through Neural Architecture Search (DiNTS) that processes 96x96x96 pixel patches from CT scans to segment eight abdominal organs and structures. The model achieves a mean Dice score of 0.88 across all structures, including liver, spleen, pancreas, stomach, gallbladder, and vascular structures (artery and portal vein). multi\_organ\_segmentation - ## Pancreas and Tumor DiNTS Segmentation (v0.5.2) A 3D segmentation model optimized through Neural Architecture Search (DiNTS) that processes 96x96x96 pixel patches from CT scans to segment pancreas and pancreatic tumors. The model architecture was automatically discovered to balance accuracy and computational efficiency, achieving a mean Dice score of 0.62 across both structures. pancreas\_ct\_dints\_segmentation - ## Pathology Nuclei Classification (v0.2.2) A deep learning model based on the HoVer-Net architecture that classifies nuclei in H&E-stained histology images. The model processes 128x128 pixel RGB images with nuclei masks and classifies four distinct cell types: inflammatory, epithelial, spindle-shaped, and other nuclei pathology\_nuclei\_classification - ## Pathology NuClick Annotation (v0.2.3) An interactive nuclei segmentation model based on the NuClick framework. The model processes 128x128 pixel RGB images with positive and negative click signals to generate nuclei segmentation masks. Trained on the CoNSeP dataset pathology\_nuclick\_annotation - ## Pathology Tumor Detection (v0.6.4) A deep learning model for detecting metastatic tissue in whole-slide pathology images. The model processes 224x224 pixel RGB patches and provides probability scores for metastasis detection. Trained on the Camelyon16 dataset pathology\_tumor\_detection - ## Pediatric Abdominal CT Segmentation (v0.4.6) A 3D segmentation model for liver, spleen, and pancreas in pediatric abdominal CT images. The model processes 96x96x96 pixel patches and provides segmentation masks. Pre-trained on TotalSegmentator, TCIA and BTCV datasets and fine-tuned on Cincinnati Children's Healthy Pediatric Dataset. pediatric\_abdominal\_ct\_segmentation - ## Prostate MRI Anatomy (v0.3.6) A 3D segmentation model that differentiates between central gland and peripheral zone within the prostate in MRI images. The model processes 96x96x96 pixel patches and provides segmentation masks. prostate\_mri\_anatomy - ## Renal Structures CECT Segmentation (v0.2.3) A 3D UNet-based segmentation model for comprehensive renal structure analysis in contrast-enhanced CT scans. The model processes 96x96x96 voxel patches and identifies six anatomical structures: arteries, veins, ureters, parenchyma, cysts, and tumors. renalStructures\_CECT\_segmentation - ## Renal Structures UNEST Segmentation (v0.2.7) A transformer-based 3D segmentation model that delineates kidney cortex, medulla, and pelvicalyceal system in CT images. The model processes 96x96x96 pixel patches and provides segmentation masks for detailed morphological analysis. renalStructures\_UNEST\_segmentation - ## retinalOCT\_RPD\_segmentation (v0.0.1) This network detects and segments Reticular Pseudodrusen (RPD) instances in Optical Coherence Tomography (OCT) B-scans which can be presented in a vol or dicom format. retinalOCT\_RPD\_segmentation - ## Spleen CT Segmentation (v0.6.1) A 3D segmentation model for spleen delineation in CT images. The model processes 96x96x96 pixel patches and provides segmentation masks for spleen tissue. Trained on the Medical Segmentation Decathlon dataset. spleen\_ct\_segmentation - ## Spleen DeepEdit Interactive Segmentation (v0.5.8) An interactive 3D segmentation model that processes 128x128x128 pixel patches from CT scans to segment the spleen. The model incorporates user-provided point annotations through the DeepEdit framework. It accepts positive and negative click inputs to refine segmentation boundaries in real-time. spleen\_deepedit\_annotation - ## Swin UNETR BTCV Multi-organ Segmentation (v0.5.8) A 3D segmentation model based on the Swin UNETR architecture that processes 96x96x96 pixel patches from CT scans to segment 13 abdominal organs and structures. The model utilizes self-supervised pre-training and hierarchical transformer blocks. swin\_unetr\_btcv\_segmentation - ## Valve Landmarks Regression (v0.5.2) A cardiac valve landmark detection model that localizes 10 valve insertion points throughout the cardiac cycle in long-axis MR images. The model processes 256x256 pixel images and outputs 2D coordinates for mitral, aortic, and tricuspid valve insertion points, enabling 3D finite element modeling for cardiac simulation. valve\_landmarks - ## Ventricular Short Axis 3-Label Segmentation (v0.3.5) A cardiac MRI segmentation model that delineates three key structures in 2D short-axis images: left ventricle blood pool, myocardium, and right ventricle blood pool. The model processes 256x256 pixel images and provides segmentation masks for functional assessment of cardiac structures throughout the cardiac cycle. ventricular\_short\_axis\_3label - ## VISTA-2D: Cell Instance Segmentation (v0.4.0) VISTA-2D is a flow-based cell instance segmentation model for microscopy images. It processes 256x256 RGB images and generates instance masks with unique labels for each cell. The model supports brightfield, fluorescence, and phase contrast imaging, handling touching cells and overlapping instances. vista2d - ## VISTA-3D: Versatile Imaging SegmenTation and Annotation (v0.5.11) A 3D segmentation model that processes 128x128x128 pixel patches from CT scans to identify and delineate over 130 anatomical structures. The model employs zero-shot learning capabilities to adapt to new anatomical targets without retraining, supporting comprehensive volumetric analysis of organs, bones, muscles, and pathological findings. vista3d - ## Whole Body CT Segmentation (v0.2.7) A SegResNet-based volumetric segmentation model that segments 104 distinct anatomical structures from CT scans. The model processes 96x96x96 pixel patches and provides segmentation masks for major organs, bones, muscles, and vascular structures throughout the body, trained on TotalSegmentator data. wholeBody\_ct\_segmentation - ## Whole Brain Large UNEST Segmentation (v0.2.7) A transformer-based 3D segmentation model that identifies 133 distinct brain structures in T1W MRI scans. The model processes 96x96x96 pixel patches and provides segmentation masks for comprehensive neuroanatomical analysis. wholeBrainSeg\_Large\_UNEST\_segmentation --- --- title: "MONAI - Medical AI Skills" description: "Verified, agent-callable skills for medical imaging: DICOM utilities, segmentation, generation, and reasoning built on MONAI workflows and NVIDIA MedTech models." canonical: https://project-monai.github.io/skills.html audience: [engineer, researcher] last_updated: 2026-09-08 source: skills.html --- Agentic AI # Medical AI Skills for Agentic Workflows A verified catalog of agent-callable skills for medical imaging: DICOM utilities, CT and MR segmentation, synthetic data generation, and image reasoning, all built on MONAI workflows and NVIDIA MedTech models. [Browse Skills](#skills) [View on GitHub](https://github.com/NVIDIA-Medtech/medical-AI-skills) 12 Skills 9 Paired Verifiers Agent Skills Spec Apache-2.0 Overview ## Trusted Building Blocks for Medical AI Agents Each skill wraps one medical AI tool through its documented entry point, so agents and engineers can discover, invoke, chain, and reproduce it in their own environments. Skills are published only after passing verification and evaluation through a domain-aware evaluation engine. ### Agent-Callable Every skill ships a `SKILL.md` following the open Agent Skills specification, a machine-readable manifest, and scripts that emit structured JSON, ready for Claude Code and other agents. ### Verified & Evidenced A skill can exit successfully and still produce output you cannot trust. Manifests encode medtech invariants, every run yields a reproducible evidence pack, and paired verifiers audit domain quality. ### Ecosystem-Native Skills wrap MONAI-based workflows, MONAI bundles, DICOM utilities, and NVIDIA MedTech models, bringing the medical imaging ecosystem you already use into agentic pipelines. Get Started ## Add a Skill to Your Agent Install any skill straight from the catalog with a single command; the CLI copies the skill into your agent's expected location. For the full trust harness with evidence packs, paired verifiers, and the gate ladder, clone the repository and use the Makefile targets. - Pick a skill and target agent interactively, or pass flags for CI - Your agent reads SKILL.md to learn when and how to run the tool - Skills emit structured JSON, so outputs chain cleanly into the next step [Read the usage guide](https://github.com/NVIDIA-Medtech/medical-AI-skills/blob/dev/docs/using-skills.md) ``` # Interactive: pick a skill, pick an agent npx skills add NVIDIA-Medtech/medical-AI-skills # Non-interactive, e.g. for Claude Code npx skills add NVIDIA-Medtech/medical-AI-skills \ --skill nv-segment-ct \ --agent claude-code --yes # Full trust harness: clone and run with evidence git clone https://github.com/NVIDIA-Medtech/medical-AI-skills make run-skill SKILL=nv-segment-ct ``` Skills follow the open [agentskills.io specification](https://agentskills.io/specification), so they work with any compliant agent, not just one vendor's. Catalog ## Available Skills From GPU-free DICOM preflight to whole-body segmentation and synthetic data generation. Each card links to the skill's folder with its SKILL.md, manifest, and fixtures. [ DICOM CPU ### dicom-series-preflight GPU-free preflight of a DICOM series folder: corruption, orientation, PHI-tag presence, and series consistency checks before conversion or inference. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/skills/dicom-series-preflight)[ DICOM CPU ### dicom-metadata-extract Extract selected metadata from a DICOM file and flag standard-tag PHI presence. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/skills/dicom-metadata-extract)[ DICOM CPU ### dicom-series-to-volume Convert a CT DICOM series folder to an HU-calibrated NIfTI volume with affine evidence. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/skills/dicom-series-to-volume)[ Segmentation GPU ### nv-segment-ct Run NV-Segment-CT (VISTA3D) whole-body segmentation on CT NIfTI volumes with label-map evidence. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/skills/nv-segment-ct)[ Segmentation GPU ### nv-segment-ctmr Run NV-Segment-CTMR segmentation on CT or MRI NIfTI volumes with label-map evidence. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/skills/nv-segment-ctmr)[ Segmentation GPU ### nv-segment-ct-finetune Auto-configuring VISTA3D continual-learning finetune on CT NIfTI labels via monai.bundle run. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/skills/nv-segment-ct-finetune)[ Generation GPU ### nv-generate-ct-rflow Generate synthetic CT volumes with paired 132-class masks using NV-Generate-CTMR rectified-flow synthesis. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/skills/nv-generate-ct-rflow)[ Generation GPU ### nv-generate-mr Generate synthetic body MRI volumes with NV-Generate-CTMR rflow-mr. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/skills/nv-generate-mr)[ Generation GPU ### nv-generate-mr-brain Generate synthetic brain MRI volumes with NV-Generate-CTMR rflow-mr-brain. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/skills/nv-generate-mr-brain)[ Generation GPU ### nv-generate-mr-brain-finetune Finetune the NV-Generate-CTMR MR-brain diffusion UNet from a NIfTI datalist. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/skills/nv-generate-mr-brain-finetune)[ Generation GPU ### nv-generate-vae-finetune Finetune the NV-Generate-CTMR MAISI VAE from CT or MRI NIfTI datalists. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/skills/nv-generate-vae-finetune)[ Reasoning GPU ### nv-reason-cxr Chest X-ray reasoning with NV-Reason-CXR-3B on a user-provided PNG or JPEG image. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/skills/nv-reason-cxr) [View the Full Skill Index](https://github.com/NVIDIA-Medtech/medical-AI-skills/blob/dev/SKILL_INDEX.md) Community ## Get Involved The catalog is Apache-2.0 and open to contributions from across the MONAI ecosystem: wrap a tool you rely on, improve the docs, or strengthen the evaluation harness. Because skills follow an open specification, you can also publish your own skill catalogs for the community. [ ### Contribute a Skill Wrap a new tool, add a paired verifier, or extend the harness. Contribution lanes, required proof, and DCO sign-off are all documented. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/blob/dev/CONTRIBUTING.md)[ ### Documentation Guides for using and authoring skills, trust and evidence, agent task maps, and skill-vs-readme comparison studies. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/docs)[ ### Eval Engine The evidence-pack harness: a 10-step gate ladder, reproducible runs with environment locks, drift comparison, and replay. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/eval_engine)[ ### Verifiers & Benchmarks Skill-shaped auditors that check domain invariants like anatomy plausibility, geometry, HU ranges, and PHI scope, plus dataset benchmark manifests. ](https://github.com/NVIDIA-Medtech/medical-AI-skills/tree/dev/verifiers) Medical AI Skills is maintained by the NVIDIA MedTech team under the Apache-2.0 license. It is a tools catalog for engineering use, not a clinical, diagnostic, or regulatory tool, and generated outputs are engineering artifacts, not clinical endorsements. Downloaded model weights carry their own licenses. --- --- title: "Tutorials & Learning Resources | MONAI" description: "Hands-on tutorials, notebooks, and learning resources for MONAI medical imaging AI framework." canonical: https://project-monai.github.io/tutorials.html audience: [researcher, engineer, newcomer] last_updated: 2026-09-08 source: tutorials.html --- Learn # Tutorials & Learning Resources Hands-on guides to help you master medical AI with MONAI. [Browse Tutorials](#tutorials) Framework All Core Label Deploy Level All Beginner Intermediate Advanced Tutorials ## Hands-On Learning Curated notebooks, documentation, and video resources for learning MONAI. [ Core Beginner ### Getting Started with MONAI Introduction to MONAI Core concepts and basic operations. Notebook ](https://github.com/Project-MONAI/tutorials/blob/main/2d_classification/mednist_tutorial.ipynb)[ Core Beginner ### Spleen 3D Segmentation End-to-end 3D medical image segmentation tutorial. Notebook ](https://github.com/Project-MONAI/tutorials/blob/main/3d_segmentation/spleen_segmentation_3d.ipynb)[ Core Beginner ### Transforms and Datasets Learn MONAI's data loading and transform pipeline. Notebook ](https://github.com/Project-MONAI/tutorials/blob/main/modules/transforms_demo_and_double.ipynb)[ Core Intermediate ### Auto3DSeg Tutorial Automated 3D segmentation pipeline with algorithm selection. Notebook ](https://github.com/Project-MONAI/tutorials/tree/main/auto3dseg)[ Core Intermediate ### SwinUNETR Segmentation Swin Transformer-based multi-organ segmentation on BTCV. Notebook ](https://github.com/Project-MONAI/tutorials/blob/main/3d_segmentation/swin_unetr_btcv_segmentation_3d.ipynb)[ Core Intermediate ### Custom Transforms Creating your own MONAI transforms. Notebook ](https://github.com/Project-MONAI/tutorials/blob/main/modules/custom_transforms.ipynb)[ Core Advanced ### MAISI - Medical Image Synthesis Generate synthetic 3D CT images with latent diffusion. Notebook ](https://github.com/Project-MONAI/tutorials/tree/main/generation/maisi)[ Core Advanced ### VISTA-3D Interactive Segmentation Interactive 3D segmentation with point prompts. Notebook ](https://github.com/Project-MONAI/tutorials/tree/main/vista3d)[ Core Advanced ### Federated Learning Privacy-preserving distributed training with MONAI. Notebook ](https://github.com/Project-MONAI/tutorials/tree/main/federated_learning)[ Label Beginner ### Getting Started with MONAI Label Setup and first annotation session. Docs ](https://docs.monai.io/projects/label/en/latest/quickstart.html)[ Label Beginner ### 3D Slicer Integration Connect MONAI Label with 3D Slicer. Docs ](https://docs.monai.io/projects/label/en/latest/installation.html)[ Label Intermediate ### Active Learning Strategies Implement active learning for efficient annotation. Notebook ](https://github.com/Project-MONAI/tutorials/tree/main/monailabel)[ Label Intermediate ### Custom MONAI Label App Build your own annotation application. Docs ](https://docs.monai.io/projects/label/en/latest/appdeployment.html)[ Deploy Beginner ### Creating Your First MAP Package a model as a MONAI Application Package. Docs ](https://docs.monai.io/projects/monai-deploy-app-sdk/en/stable/getting_started/index.html)[ Deploy Beginner ### DICOM Data Processing Process DICOM data in deployment pipelines. Docs ](https://docs.monai.io/projects/monai-deploy-app-sdk/en/stable/)[ All Frameworks Beginner ### MONAI Bootcamp 2023 Comprehensive workshop series covering Core, Label, and Deploy. Video ](https://www.youtube.com/playlist?list=PLtoSVSQ2XzyD4zuSmTtgIciF0ZjzIDr7Y)[ All Frameworks Intermediate ### MONAI at MICCAI Conference workshops and presentations. Video ](https://www.youtube.com/c/Project-MONAI) Resources ## Explore More Dive deeper with the full range of MONAI learning materials. [ ### Tutorials Repository Full collection of Jupyter notebooks on GitHub. ](https://github.com/Project-MONAI/tutorials)[ ### Official Documentation Comprehensive API reference and guides. ](https://monai.readthedocs.io/en/stable/)[ ### YouTube Channel Video tutorials, bootcamps, and talks. ](https://www.youtube.com/c/Project-MONAI)[ ### Bootcamp Materials Complete MONAI Bootcamp workshop series. ](https://www.youtube.com/playlist?list=PLtoSVSQ2XzyD4zuSmTtgIciF0ZjzIDr7Y) ## Start Building with MONAI Ready to build your first medical AI application? Get started with installation and explore pre-trained models. [Get Started](core.html#quick-start) [Browse Model Zoo](model-zoo.html) --- --- title: "MONAI - About Us" description: "Project MONAI was started by NVIDIA and King's College London to build an inclusive community of AI researchers developing and exchanging best practices for AI in healthcare imaging." canonical: https://project-monai.github.io/about.html audience: [engineer] last_updated: 2026-09-08 source: about.html --- About # The community behind medical imaging AI Project MONAI brings together researchers, clinicians, and industry engineers to advance medical-imaging AI. Started by NVIDIA and King's College London, it has grown into a community of 30+ institutions developing and sharing best practices, tools, and frameworks across academia and enterprise. Our mission is wider adoption of AI in healthcare imaging through validated, open tools. Building on the foundations of NVIDIA Clara Train, NiftyNet, DLTK, and DeepNeuro, MONAI Core is now the community's flagship framework, and every contribution is reviewed and merged in the open. Leadership ## Advisory Board The Advisory Board sets MONAI's strategic direction. Its members, leaders from academia, healthcare, and industry, provide oversight across technical development, clinical validation, education, and community engagement, and keep the project aligned with what researchers and clinicians actually need. Our [Working Groups](working-groups.html) are driven at the Advisory Board level: each working-group lead sits on the board. These teams focus on core development, deployment, outreach and education, benchmarking, federated learning, AI-human interaction, and clinical applications such as ophthalmology and ultrasound. ### Advisory Board Members ![Aaron Lee](/assets/img/people/lee-aaron.png) #### Aaron Lee Ophthalmology Working Group Chair ![Annika Reinke](/assets/img/people/annika-reinke.jpg) #### Annika Reinke Benchmark Working Group Chair ![Ben Murray](/assets/img/people/ben-murray.png) #### Ben Murray Thought Leader ![Bruce Hashemian](/assets/img/people/bruce-hashemian.jpeg) #### Bruce Hashemian AI-Human Interaction Chair & Thought Leader ![Carole Sudre](/assets/img/people/carole-sudre.jpg) #### Carole Sudre Benchmark Working Group Chair ![David Bericat](/assets/img/people/david-bericat.jpg) #### David Bericat Deploy Working Group Chair ![Eric Kerfoot](/assets/img/people/eric-kerfoot.png) #### Eric Kerfoot Core Developers & Outreach, Adoption, Education Chair ![Haris Shuaib](/assets/img/people/haris-shuaib.jpg) #### Haris Shuaib Deploy Working Group Chair ![Holger Roth](/assets/img/people/holger-roth.jpg) #### Holger Roth Federated Learning Chair ![Jayashree Kalpathy-Cramer](/assets/img/people/jayashree-kalpathy-cramer.jpg) #### Jayashree Kalpathy-Cramer Ophthalmology Working Group Chair ![Jorge Cardoso](/assets/img/people/jorge-cardoso.jpg) #### Jorge Cardoso Arbitration Board Chair & Thought Leader ![Justin Kirby](/assets/img/people/justin-kirby.png) #### Justin Kirby Thought Leader ![Lena Maier-Hein](/assets/img/people/lena-maier-hein.jpg) #### Lena Maier-Hein Thought Leader ![Mahdi Azizian](/assets/img/people/mahdi-azizian.jpg) #### Mahdi Azizian Arbitration Board Chair & Thought Leader ![Michael Baumgartner](/assets/img/people/michael-baumgartner.jpg) #### Michael Baumgartner Thought Leader ![Michael Zephyr](/assets/img/people/michael-zephyr.jpg) #### Michael Zephyr Outreach, Adoption, and Education Chair ![Nic Ma](/assets/img/people/nic-ma.jpg) #### Nic Ma Core Developers Chair ![Prerna Dogra](/assets/img/people/prerna-dogra.png) #### Prerna Dogra Thought Leader ![Rafael Garcia-Dias](/assets/img/people/rafael-garcia-dias.jpg) #### Rafael Garcia-Dias Federated Learning Chair ![Ralf Floca](/assets/img/people/ralf-floca.png) #### Ralf Floca AI-Human Interaction Chair & Arbitration Board ![Saad Nadeem](/assets/img/people/saad-nadeem.png) #### Saad Nadeem AI-Human Interaction Chair ![Sebastien Ourselin](/assets/img/people/sebastien-ourselin.png) #### Sebastien Ourselin Thought Leader ![Barbaros Selnur Erdal](/assets/img/people/barbaros-selnur-erdal.jpg) #### Barbaros Selnur Erdal Deploy Working Group Chair ![Stephen Aylward](/assets/img/people/stephen-aylward.png) #### Stephen Aylward Chair of the Advisory Board ![Supriya Thathachary](/assets/img/people/supriya-thathachary.jpg) #### Supriya Thathachary Outreach, Adoption, and Education Chair ![Yun Liu](/assets/img/people/yun-liu.jpeg) #### Yun Liu Core Developers Chair Alumni ## Emeritus Members We recognize the founding members who shaped MONAI's earliest direction, along with the board and working-group leads who have since stepped back. Their work continues to shape the project and the field. ![Daniel Rubin](/assets/img/people/daniel-rubin.png) #### Daniel Rubin Founding Member ![Keyvan Farahani](/assets/img/people/keyvan-farahani.jpg) #### Keyvan Farahani Founding Member ![Klaus Maier-Hein](/assets/img/people/klaus-maier-hein.jpg) #### Klaus Maier-Hein Founding Member ![Nassir Navab](/assets/img/people/nassir-navab.png) #### Nassir Navab Founding Member ![Nasir Rajpoot](/assets/img/people/nasir-rajpoot.png) #### Nasir Rajpoot Founding Member ![S. Kevin Zhou](/assets/img/people/kevin-zhou.png) #### S. Kevin Zhou Founding Member ![Andy Feng](/assets/img/people/andy-feng.jpeg) #### Andy Feng Founding Member ![Wenqi Li](/assets/img/people/wenqi-li.jpg) #### Wenqi Li Founding Member ![Sai Praneeth Karimireddy](/assets/img/people/sai-praneeth-karimireddy.jpg) #### Sai Praneeth Karimireddy Federated Learning Chair ![Marc Modat](/assets/img/people/marc-modat.jpg) #### Marc Modat Education Working Group Chair Get Involved ## Help shape what MONAI becomes next MONAI is built in the open: code, governance, and roadmap. Join a working group, contribute a pull request, or bring a clinical use case to the community. [Join the Community](https://github.com/Project-MONAI/MONAI/blob/master/CONTRIBUTING.md) [Explore Working Groups](working-groups.html) --- --- title: "MONAI - Working Groups" description: "MONAI Working Groups are specialized teams that drive innovation and progress in specific areas of medical imaging AI." canonical: https://project-monai.github.io/working-groups.html audience: [engineer] last_updated: 2026-09-08 source: working-groups.html --- Community # Working Groups Small groups of researchers and engineers who own a specific area of MONAI: ultrasound, federated learning, ophthalmology, deploy, human-AI interaction. Each group meets regularly, ships code and publications, and reports to the Advisory Board. Each working group is led by chairs who serve as members of the MONAI [Advisory Board](about.html#advisory-board), which keeps each group aligned with MONAI's overall mission. Navigate ## Find Your Group Jump directly to a working group or browse through all groups below to learn about their missions and leadership. [ Deploy Research to clinical production ](#deploy)[ Developers Committee Technical excellence & code quality ](#developers)[ Evaluation & Benchmarking Guidelines & reproducibility tools ](#evaluation)[ Federated Learning Privacy-preserving collaboration ](#federated)[ Human-AI Interaction Interactive AI workflows ](#human-ai)[ Ophthalmology AI for eye disease diagnosis ](#ophthalmology)[ Outreach, Adoption & Education Training materials & community growth ](#outreach)[ Ultrasound Scalable ultrasound AI pipelines ](#ultrasound) Our Teams ## All Working Groups From technical implementation to clinical applications, our working groups ensure that MONAI remains at the forefront of healthcare AI innovation. ### Deploy Working Group This working group aims to define how to close the existing gap from research and development to clinical production environments by bringing AI models into medical applications and clinical workflows with the end goal of helping improve patient care. The focus includes defining the open high-level functional architecture and determining which components and standard APIs are required. By collaborating with the MONAI developers, the group will move from requirements to implemented solutions. Group Leads ![Barbaros Selnur Erdal](/assets/img/people/barbaros-selnur-erdal.jpg) Barbaros Selnur Erdal Mayo Clinic ![David Bericat](/assets/img/people/david-bericat.jpg) David Bericat NVIDIA ![Haris Shuaib](/assets/img/people/haris-shuaib.jpg) Haris Shuaib Guy's and St Thomas' NHS Foundation Trust [Learn More](wg_deploy.html) ### Developers Committee Working Group This working group aims to establish and maintain technical excellence in MONAI Core by coordinating development efforts and ensuring high code quality standards. The focus includes overseeing architectural decisions, establishing development guidelines, and maintaining technical documentation with the end goal of creating a sustainable framework for medical imaging AI research and development. By fostering collaboration between contributors, the group ensures consistent implementation of best practices across the codebase. Group Leads ![Nic Ma](/assets/img/people/nic-ma.jpg) Nic Ma NVIDIA ![Eric Kerfoot](/assets/img/people/eric-kerfoot.png) Eric Kerfoot King's College London ![Yun Liu](/assets/img/people/yun-liu.jpeg) Yun Liu NVIDIA [Learn More](wg_developers.html) ### Evaluation and Benchmarking Working Group The Evaluation and Benchmarking MONAI working group aims at providing guidelines, infrastructure, and practical tools for evaluation and benchmarking of medical image analysis methods. It focuses on leading the community towards the identification and adoption of best practices for evaluation and benchmarking and on identifying practical solutions to improve reproducibility. Group Leads ![Annika Reinke](/assets/img/people/annika-reinke.jpg) Annika Reinke DKFZ ![Carole Sudre](/assets/img/people/carole-sudre.jpg) Carole Sudre UCL [Learn More](wg_evaluation_benchmark.html) ### Data Quality and Federated Learning Working Group This working group aims to advance collaborative medical AI research through secure and efficient federated learning implementations. The focus includes developing standardized workflows, ensuring data compatibility, and creating modular components with the end goal of enabling distributed learning across institutions while preserving data privacy. By establishing best practices for federated learning, the group facilitates multi-institutional collaboration in medical AI research. Group Leads ![Holger Roth](/assets/img/people/holger-roth.jpg) Holger Roth NVIDIA ![Rafael Garcia-Dias](/assets/img/people/rafael-garcia-dias.jpg) Rafael Garcia-Dias KCL [Learn More](wg_federated_learning.html) ### Human-AI Interaction Working Group This working group aims to advance, standardize, and support human-AI interaction cycles through well-defined interfaces and community-driven development. The focus includes developing standardized APIs for AI prompting, uncertainty communication, and annotation feedback processes with the end goal of creating reusable interactive AI workflows. By fostering community-driven specification and implementation across diverse domains including radiology, pathology, and surgery, the group enables efficient AI deployment across cloud, local, and HPC environments. Group Leads ![Ralf Floca](/assets/img/people/ralf-floca.png) Dr. Ralf Floca DKFZ ![Saad Nadeem](/assets/img/people/saad-nadeem.png) Dr. Saad Nadeem Memorial Sloan Kettering Cancer Center ![Bruce Hashemian](/assets/img/people/bruce-hashemian.jpeg) Bruce Hashemian NVIDIA [Learn More](wg_human_ai_interaction.html) ### Ophthalmology Working Group This working group aims to advance the application of AI in ophthalmology through specialized tools and algorithms. The focus includes developing analysis methods for various ophthalmic imaging modalities, creating annotation tools, and building predictive models with the end goal of improving diagnosis and treatment of eye diseases. By leveraging MONAI's capabilities, the group accelerates innovation in ophthalmic image analysis. Group Leads ![Jayashree Kalpathy-Cramer](/assets/img/people/jayashree-kalpathy-cramer.jpg) Jayashree Kalpathy-Cramer University of Colorado ![Aaron Lee](/assets/img/people/lee-aaron.png) Aaron Lee University of Washington [Learn More](wg_ophthalmology.html) ### Outreach, Adoption, and Education Working Group This working group formed by merging the former Outreach and Adoption Working Group with the former Education Working Group. It covers both sides of that merge: the learning materials that get people started with MONAI, and the community work that brings them in. Much of the material already exists, so most of the effort goes into reorganizing and maintaining it rather than writing from scratch. Group Leads ![Michael Zephyr](/assets/img/people/michael-zephyr.jpg) Michael Zephyr NVIDIA ![Supriya Thathachary](/assets/img/people/supriya-thathachary.jpg) Supriya Thathachary NVIDIA ![Eric Kerfoot](/assets/img/people/eric-kerfoot.png) Eric Kerfoot King's College London [Learn More](wg_outreach_adoption_education.html) ### Ultrasound Working Group This working group will focus on enabling reproducible, scalable AI development for ultrasound within the MONAI ecosystem. Key priorities include harmonizing data formats (e.g., DICOM for B-mode); supporting data streaming (e.g., for RF signals); standardizing annotation formats and labeling protocols for common ultrasound tasks; and defining reusable pipelines for training, inference, evaluation, and deployment. Short-term and long-term priorities and focal clinical applications will be determined by participants of the group. By collaborating with the MONAI developers, the group will move from requirements to implemented solutions. Group Leads ![Tina Kapur](/assets/img/people/tina-kapur.jpg) Tina Kapur Brigham and Women's Hospital, Harvard Medical School ![Stephen Aylward](/assets/img/people/stephen-aylward.jpg) Stephen Aylward NVIDIA [Learn More](wg_ultrasound.html) Get Involved ## Join a Working Group Working group members collaborate regularly through meetings, workshops, and joint projects, fostering knowledge exchange and driving continuous improvement in their specialized areas. Get started by exploring the groups above or reaching out to the community. [View on GitHub](https://github.com/Project-MONAI) [Advisory Board](about.html#advisory-board) --- --- title: "MONAI - Deploy Working Group" description: "Bridging the gap between research and clinical production environments by bringing AI models into medical applications and clinical workflows." canonical: https://project-monai.github.io/wg_deploy.html audience: [engineer] last_updated: 2026-09-08 source: wg_deploy.html --- # Deploy Working Group Mission Statement This working group aims to define how to close the existing gap from research and development to clinical production environments by bringing AI models into medical applications and clinical workflows with the end goal of helping improve patient care. The focus includes defining the open high-level functional architecture and determining which components and standard APIs are required. By collaborating with the MONAI developers, the group will move from requirements to implemented solutions. ## Initiatives ### Current Projects - • Development of the MONAI Deploy SDK for model packaging - • Implementation of standardized deployment workflows - • Creation of reference architectures for clinical deployment - • Integration with existing hospital PACS and RIS systems ### Upcoming Focus Areas - • Enhanced security and compliance frameworks - • Performance optimization for edge deployment - • EHR integration standards and protocols - • Multi-site deployment orchestration ## Group Leads ![Barbaros Selnur Erdal](/assets/img/people/barbaros-selnur-erdal.jpg) Barbaros Selnur Erdal Technical Director, Center for Augmented Intelligence in Imaging Mayo Clinic Deploy Working Group Chair [View Profile](https://www.linkedin.com/in/barbaros-selnur-erdal-1690055/) ![David Bericat](/assets/img/people/david-bericat.jpg) David Bericat Healthcare AI Solutions NVIDIA Deploy Working Group Chair [View Profile](https://www.linkedin.com/in/davidbericat/) ![Haris Shuaib](/assets/img/people/haris-shuaib.jpg) Haris Shuaib Chief Executive Officer Newton's Tree Deploy Working Group Chair [View Profile](https://www.linkedin.com/in/haris-shuaib-91519b55/) ## Resources ### Development Resources - • [MONAI Deploy GitHub Repository](https://github.com/Project-MONAI/monai-deploy) - • [MONAI Deploy App SDK Documentation](https://monai.readthedocs.io/projects/monai-deploy-app-sdk/en/latest/) ## Collaboration Opportunities ### Development Contributions - • Contribute to the development of deployment solutions through our [GitHub repository](https://github.com/Project-MONAI/monai-deploy) - • Share your clinical deployment experience and requirements - • Participate in defining standards and best practices ### Research & Standards - • Join our regular working group meetings - • Contribute to deployment best practices - • Help shape the future of medical AI deployment --- --- title: "MONAI - Developers Committee Working Group" description: "Coordinating the development and engineering alignment of MONAI Core." canonical: https://project-monai.github.io/wg_developers.html audience: [engineer] last_updated: 2026-09-08 source: wg_developers.html --- # Developers Committee Working Group Mission Statement The Developers Committee Working Group serves as the central coordination point for MONAI Core development, ensuring engineering excellence, maintaining code quality, and driving technical decisions. This committee oversees the architectural evolution of MONAI, establishes development standards, and coordinates engineering efforts across the project. ## Initiatives ### Technical Excellence - • Architecture decisions and standards - • Code review and quality assurance - • Performance optimization strategies - • Technical debt management ### Development & Standards - • Release planning and coordination - • Testing and CI/CD practices - • Security and compatibility standards - • Documentation and API guidelines ## Group Leads ![Nic Ma](/assets/img/people/nic-ma.jpg) Nic Ma Senior Engineering Manager NVIDIA Core Developers Chair [View Profile](https://www.linkedin.com/in/nic-ma-83b201103) ![Eric Kerfoot](/assets/img/people/eric-kerfoot.png) Eric Kerfoot Software Architect in Medical Engineering King's College London Core Developers & Education Chair [View Profile](https://kclpure.kcl.ac.uk/portal/eric.kerfoot.html) ![Yun Liu](/assets/img/people/yun-liu.jpeg) Yun Liu Software Engineer NVIDIA Core Developers Chair [View Profile](https://www.linkedin.com/in/yun-liu-cn-221337245) ## Resources ### Development Resources - • [MONAI Core Repository](https://github.com/Project-MONAI/MONAI) - • [Contribution Guidelines](https://monai.readthedocs.io/en/latest/contributing.html) - • [Issue Tracker](https://github.com/Project-MONAI/MONAI/issues) ### Community Resources - • [Developer Discussions](https://github.com/Project-MONAI/MONAI/discussions) - • [Technical Documentation](https://monai.readthedocs.io) - • [Developer Wiki](https://github.com/Project-MONAI/MONAI/wiki) ## Collaboration Opportunities ### Development Contributions - • Contribute to MONAI Core development - • Review and improve code quality - • Enhance test coverage and documentation ### Community Engagement - • Join our regular working group meetings - • Participate in technical discussions - • Share expertise and best practices --- --- title: "MONAI - Evaluation and Benchmarking Working Group" description: "Providing guidelines, infrastructure, and practical tools for quality-controlled validation and benchmarking of medical image analysis methods." canonical: https://project-monai.github.io/wg_evaluation.html audience: [engineer] last_updated: 2026-09-08 source: wg_evaluation_benchmark.html --- # Evaluation and Benchmarking Working Group Mission Statement The Evaluation and Benchmarking MONAI working group aims at providing guidelines, infrastructure, and practical tools for evaluation and benchmarking of medical image analysis methods. It focuses on leading the community towards the identification and adoption of best practices for evaluation and benchmarking and on identifying practical solutions to improve reproducibility. ## Highlights ### Recommendations - • [Metrics Reloaded: recommendations for image analysis validation](https://www.nature.com/articles/s41592-023-02151-z) - • [Understanding metric-related pitfalls in image analysis validation](https://www.nature.com/articles/s41592-023-02150-0) - • [Metrics Reloaded Toolkit](https://metrics-reloaded.dkfz.de/) ### Implementation of recommendations - • [MONAI Evaluation Metrics](https://github.com/Project-MONAI/MetricsReloaded/) - • [Metrics Documentation](https://monai.readthedocs.io/en/latest/metrics.html) ### Related resources - • [Biomedical Image Analysis ChallengeS (BIAS) Initiative](https://www.dkfz.de/en/imsy/research/biomedical-image-analysis-challenges-bias-initiative) - • [Rankings Reloaded](https://www.rankings-reloaded.de/) ## Group Leads ![Dr. Annika Reinke](/assets/img/people/annika-reinke.jpg) Dr. Annika Reinke Deputy Head and Group Lead Validation of Intelligent Systems German Cancer Research Center (DKFZ) Benchmark Working Group Chair [View Profile](https://www.dkfz.de/en/employees/annika-reinke) ![Dr. Carole Sudre](/assets/img/people/carole-sudre.jpg) Dr. Carole Sudre Associate Professor University College London Benchmark Working Group Chair [View Profile](https://profiles.ucl.ac.uk/39648-carole-sudre) ## Meeting Notes ### GitHub Wiki - • [Access all meeting notes](https://github.com/Project-MONAI/MONAI/wiki/Evaluation-and-Benchmarks-Working-Group-Meeting-Notes) ## Ongoing Projects ### Reporting Guidelines Taskforce (Lead - Olivier Colliot) - • Surveying current reporting practices and identifying areas for improvement - • Development of guidelines around results reporting with a focus on statistical aspects - • Identification of proper calculation and methods for various procedures (e.g., confidence intervals) across different tasks and validation metrics - • Implementation of recommended calculations for MONAI users ### Benchmarking Datasets Taskforce (Lead - Michela Antonelli) - • Data quality review for MICCAI 2025 lighthouse challenges - • Identification of key characteristics for benchmarking datasets - • Encouragement to develop new datasets according to best practice - • Identification of relevant historical datasets to be used for benchmarking - • Implementation of guidelines for upcoming datasets ## Collaboration Opportunities ### Community Engagement - • Join our regular surveys - • Contribute to evaluation metrics testing - • Share your expertise in validation and benchmarking - • Participate in standards development --- --- title: "MONAI - Federated Learning Working Group" description: "Accelerating medical research and clinical translation through federated learning in medical imaging." canonical: https://project-monai.github.io/wg_federated_learning.html audience: [engineer] last_updated: 2026-09-08 source: wg_federated_learning.html --- # Data Quality and Federated Learning Working Group Mission Statement This working group aims to advance collaborative medical AI research through secure and efficient federated learning implementations. The focus includes developing standardized workflows, ensuring data compatibility, and creating modular components with the end goal of enabling distributed learning across institutions while preserving data privacy and reliability. By fostering a community of academic, clinical, and industry experts, the group establishes best practices and innovative solutions that enhance the accessibility and impact of federated learning in medical imaging. ## Initiatives ### Deliverables Development - • Database integration with federated query capabilities (XNAT, Flywheel) - • Modular FL components and standardized workflows - • Privacy-preserving validation methods and data quality metrics - • Benchmarking suites and performance evaluation - • Cohort creation and management tools ### Community & Research - • Best practices and guidelines for trustworthy FL - • Documentation and educational resources - • FL challenges and workshops at major conferences - • Research publications in NeurIPS FL and MICCAI - • Cross-institutional collaboration and knowledge sharing ## Group Leads ![Dr. Holger Roth](/assets/img/people/holger-roth.jpg) Dr. Holger Roth Principal Research Scientist NVIDIA Federated Learning Chair [View Profile](https://research.nvidia.com/person/holger-roth/) ![Dr. Rafael Garcia-Dias](/assets/img/people/rafael-garcia-dias.jpg) Dr. Rafael Garcia-Dias Senior AI Engineer KCL Federated Learning Chair [View Profile](https://www.linkedin.com/in/garcia-dias/) ## Resources ### Development Resources - • [Federated Learning Documentation](https://monai.readthedocs.io/en/latest/fl.html) - • [FL Tutorials](https://github.com/Project-MONAI/tutorials/tree/main/federated_learning) ## Meeting Recordings ### YouTube Playlist Watch all our working group meetings and discussions on our YouTube playlist. Subscribe to stay updated with the latest developments in federated learning. [View Full Playlist](https://www.youtube.com/playlist?list=PLtoSVSQ2XzyAKQa4AkoQEviFn9pN9OZY8) ## Collaboration Opportunities ### Development Contributions - • Contribute to MONAI FL through our [GitHub repositories](https://github.com/Project-MONAI/) - • Develop new FL components and integrations - • Create example applications and use cases - • Improve documentation and tutorials ### Research & Standards - • Share expertise in federated learning - • Participate in benchmark development - • Join our regular working group meetings - • Contribute to best practice guidelines --- --- title: "MONAI - HAI Working Group: How We Work" description: "Practical operations guide for the MONAI Human-AI Interaction Working Group - principles, processes, and collaboration guidelines." canonical: https://project-monai.github.io/wg_hai_how_we_work.html audience: [engineer] last_updated: 2026-09-08 source: wg_hai_how_we_work.html --- [← Back to HAI Working Group](wg_human_ai_interaction.html) # How We Work This document describes the practical operation of the MONAI Human-AI Interaction Working Group (WG). Foundation ## Guiding Principles ### Issue-driven GitHub issues are our system of record ### Async-first Work happens primarily between meetings ### Low barrier Early and imperfect ideas are welcome ### Transparent Priorities, discussions, and agendas are public ### Community-led Anyone can contribute or claim work Culture ## The HAI Way Principles for productive collaboration, inspired by those who came before us: - • **Joining a ticket is better than starting a new one.** - • **Starting a new one is better than staying silent.** _(When in doubt, create it.)_ - • **Decisions live in issues, not in meetings.** - • **Meeting notes capture "what happened," issues capture "what we decided."** - • **Progress happens asynchronously; meetings exist to unblock.** - • **Imperfect contributions beat perfect silence.** - • **If it's important enough to discuss, it's important enough to document.** - • **Transparent is better than efficient.** _(When forced to choose.)_ - • **Anyone can contribute. Everyone can claim work.** - • **Subscribe before you criticize** - context helps. - • **If something blocks you, speak up.** We're here to unblock each other. - • **Before it's lost in Slack/Mailinglist, capture it in GitHub.** These aren't rules - they're agreements on how we work best together. Tools ## Our Coordination Hub All technical discussions, proposals, and decisions are documented in **GitHub issues**. ### WG Project Board Central coordination space for all WG activities [View Project Board](https://github.com/orgs/Project-MONAI/projects/28) ### MONAI Label Milestones Focus areas and medium-term direction (not strict deadlines) [View Milestones](https://github.com/Project-MONAI/MONAILabel/milestones) Issues ## What Belongs in an Issue? ### Issues can represent: - • Feature ideas or enhancement proposals - • Bugs or technical problems - • Discussion topics or open questions - • Workflow improvements - • Research or UX questions - • Presentation proposals for WG meetings - • Anything else relevant to human-AI interaction in medical imaging ### Important Notes **Issues don't need to be perfect.** A short description is enough to start the conversation. **Before creating a new issue:** Check if a similar one already exists. If you're unsure, create it anyway - we'd rather have duplicates than miss your input. Process ## How Work Gets Done 1 ### Claim an Issue Assign yourself and indicate you'll work on it 2 ### Work Async Progress through comments, commits, and collaboration 3 ### Others Join By subscribing, commenting, or co-developing 4 ### Share Updates Directly in the issue thread **The WG meeting is not where implementation happens.** It's where we unblock, align, and connect. Escalation ## When to Request WG Discussion If you're working on an issue and need broader input: 1. 1 Move it to or label it **"Needs Discussion"** 2. 2 A coordinator will schedule it for the next WG meeting 3. 3 It appears in the **"Next WG Meeting"** column Meetings ## WG Meeting Structure Meetings serve four purposes: ### Community touchpoint Staying connected as a group ### Unblocking discussions Addressing issues that need WG-level input ### Issue grooming Reviewing new issues, identifying duplicates, assigning milestones ### Knowledge sharing Short presentations on relevant topics (when proposed) **The "Next WG Meeting" column is the agenda.** It's frozen ~1 week before each meeting so you can decide if attendance is valuable for you. Documentation ## Meeting Documentation ### Meeting notes capture: **Organizational information only** - • Scheduling updates - • Process changes - • Announcements ### Why this approach? **All technical discussions and decisions are documented in the issues themselves**, not in meeting minutes. This ensures: - • Contributors who can't attend aren't disadvantaged - • Discussions remain searchable and accessible - • Participation works across time zones - • Context is preserved where it belongs Maintenance ## Issue Grooming Process At each meeting, we review: - • New issues created since the last meeting - • Issues without milestone assignments - • Potential duplicates - • Unclear problem statements The goal is to keep the backlog organized and actionable. Leadership ## Who Coordinates the WG? The WG is led by coordinators listed on the [WG homepage](wg_human_ai_interaction.html). ### Coordinators are responsible for: - • Scheduling and facilitating WG meetings - • Managing the "Next WG Meeting" agenda - • Issue grooming and milestone assignment - • Ensuring the process remains accessible and transparent Reach out to current coordinators if you'd like to help. Connect ## Communication Channels ### Issue-specific discussions Use the GitHub issue itself (preferred) ### Quick questions & chat - • **Mailing list:** [monai-wg-hai Google Group](https://groups.google.com/g/monai-wg-hai) - • **Slack:** [#human-ai-interaction](https://projectmonai.slack.com/archives/C0A9334ETNK) ### Join the WG [Fill out the participation form](https://docs.google.com/forms/d/e/1FAIpQLSdsgt8arOOle203KJyPvkjr3VQTyvrzzazrJumgQgdZ7voH8Q/viewform) **GitHub is the source of truth for decisions.** Discussions on Slack or the mailing list should be summarized back into relevant issues. Resolution ## Conflict Resolution When contributors disagree on an approach or decision: 1. 1 **Start with discussion** in the issue - most conflicts resolve through clarification 2. 2 **Request coordinator input** if the discussion stalls 3. 3 **Bring to WG meeting** by moving to "Needs Discussion" if broader input would help 4. 4 **Community vote** as a last resort (voting mechanism: TBD based on need) The goal is **consensus where possible, clarity always**. We value diverse perspectives and aim for solutions that serve the broadest community needs. Rationale ## Why This Approach? This structure ensures: ### Async participation works You don't need to attend meetings to contribute ### Decisions are documented Everything is searchable and traceable ### Global inclusion Contributors from all time zones can participate fully ### Scalability The WG can grow without bottlenecking on synchronous meetings Get Involved ## Getting Started 1. 1 Browse the [WG Project Board](https://github.com/orgs/Project-MONAI/projects/28) 2. 2 Look for issues that interest you 3. 3 Comment, subscribe, or claim issues you want to work on 4. 4 Create new issues for ideas not yet captured 5. 5 Join WG meetings when topics relevant to you are scheduled **Questions?** Open an issue on the project board or reach out to the [WG coordinators](wg_human_ai_interaction.html). --- --- title: "MONAI - Human-AI Interaction Working Group" description: "Standardizing and supporting Human-AI-Interaction cycles through well-defined interfaces for AIaaS." canonical: https://project-monai.github.io/wg_human_ai_interaction.html audience: [engineer] last_updated: 2026-09-08 source: wg_human_ai_interaction.html --- # Human-AI Interaction Working Group Mission Statement This working group aims to advance, standardize, and support human-AI interaction cycles through well-defined interfaces and community-driven development. The focus includes developing standardized APIs for AI prompting, uncertainty communication, and annotation feedback processes with the end goal of creating reusable interactive AI workflows. By fostering community-driven specification and implementation across diverse domains including radiology, pathology, and surgery, the group enables efficient AI deployment across cloud, local, and HPC environments. [How We Work](wg_hai_how_we_work.html) ## Get Involved ### Join Our Working Group We are looking for engaging members for the MONAI Human-AI Interaction (HAI) Working Group. If you are interested in contributing, even if it is just participating and adding to discussions at our meetings, please fill out our participation form and get in contact with us. [Working Group Participation Form](https://docs.google.com/forms/d/e/1FAIpQLSdsgt8arOOle203KJyPvkjr3VQTyvrzzazrJumgQgdZ7voH8Q/viewform) ### Meeting Information - • **Schedule:** Every 1st and 3rd Wednesday of the month - • **Time:** 4:00 pm CET / 10:00 am EST / 7:00 am PST - • **Contact:** [monai-wg-hai-contact@googlegroups.com](mailto:monai-wg-hai-contact@googlegroups.com) ## Initiatives ### Interface Development - • Standardizing AI prompting protocols and APIs - • Developing intuitive interfaces for AI result visualization - • Creating frameworks for uncertainty communication - • Implementing efficient annotation feedback loops ### Integration & Implementation - • Enhancement of MONAI Label capabilities - • Development of cross-domain applications - • Support for diverse deployment environments - • Integration with clinical workflows ## Group Leads ![Ralf Floca](/assets/img/people/ralf-floca.png) Dr. Ralf Floca Group Lead German Cancer Research Center (DKFZ) Human-AI Interaction Chair & Arbitration Board [View Profile](https://www.linkedin.com/in/ralf-floca) ![Saad Nadeem](/assets/img/people/saad-nadeem.png) Dr. Saad Nadeem Assistant Professor Memorial Sloan Kettering Cancer Center Human-AI Interaction Chair [View Profile](https://www.mskcc.org/profile/saad-nadeem) ![Bruce Hashemian](/assets/img/people/bruce-hashemian.jpeg) Dr. Bruce Hashemian AI in Healthcare NVIDIA Human-AI Interaction Chair [View Profile](https://www.linkedin.com/in/brucehashemian) ## Resources ### Development Resources - • [How We Work](wg_hai_how_we_work.html) - Our organizational process and principles - • [WG Project Board](https://github.com/orgs/Project-MONAI/projects/28) - Central coordination space - • [MONAI Label Repository](https://github.com/Project-MONAI/MONAILabel) - • [Sample Applications](https://github.com/Project-MONAI/MONAILabel/tree/main/sample-apps) ### Educational Materials - • [Latest Features & Updates](https://monai.readthedocs.io/en/latest/whatsnew.html) - • [Tutorial Repository](https://github.com/Project-MONAI/tutorials) - • [Community Discussions](https://github.com/Project-MONAI/MONAILabel/discussions) ## Collaboration Opportunities ### Development Contributions - • Contribute to MONAI Label development through our [GitHub repository](https://github.com/Project-MONAI/MONAILabel) - • Implement new interaction patterns and visualization tools - • Create sample applications and use cases ### Community Engagement - • Share your expertise in human-AI interaction design - • Participate in defining standards and best practices - • Contribute to community discussions and knowledge sharing - • [Join our working group](https://docs.google.com/forms/d/e/1FAIpQLSdsgt8arOOle203KJyPvkjr3VQTyvrzzazrJumgQgdZ7voH8Q/viewform) - meetings held 1st and 3rd Wednesday of each month --- --- title: "MONAI - Ophthalmology Working Group" description: "Developing open-source MONAI-based algorithms and tools for ophthalmology image analysis." canonical: https://project-monai.github.io/wg_ophthalmology.html audience: [engineer] last_updated: 2026-09-08 source: wg_ophthalmology.html --- # Ophthalmology Working Group Mission Statement This working group aims to advance the application of AI in ophthalmology through specialized tools and algorithms. The focus includes developing analysis methods for various ophthalmic imaging modalities, creating annotation tools, and building predictive models with the end goal of improving diagnosis and treatment of eye diseases. By leveraging MONAI's capabilities, the group accelerates innovation in ophthalmic image analysis. ## Initiatives ### Image Analysis & Processing - • Color fundus photo analysis and processing - • OCT image segmentation and classification - • Multimodal registration and fusion - • Disease classification and detection models ### Clinical Applications & Tools - • Smart annotation and segmentation tools - • Clinical workflow integration solutions - • Treatment planning support systems - • Validation and performance optimization ## Group Leads ![Dr. Jayashree Kalpathy-Cramer](/assets/img/people/jayashree-kalpathy-cramer.jpg) Dr. Jayashree Kalpathy-Cramer Chief, Division of Artificial Medical Intelligence, Ophthalmology University of Colorado Ophthalmology Working Group Chair [View Profile](https://som.cuanschutz.edu/Profiles/Faculty/Profile/36939) ![Dr. Aaron Lee](/assets/img/people/lee-aaron.png) Dr. Aaron Lee Professor University of Washington Ophthalmology Working Group Chair [View Profile](https://www.linkedin.com/in/aaron-y-lee-md-msci-a102b87/) ## Resources ### Development Resources - • [Tutorial Notebooks](https://github.com/Project-MONAI/tutorials) ## Collaboration Opportunities ### Development Contributions - • Contribute to algorithm development - • Help create analysis tools - • Develop clinical workflows ### Community Engagement - • Join our regular working group meetings - • Share your clinical expertise - • Participate in validation studies --- --- title: "MONAI - Outreach, Adoption, and Education Working Group" description: "The MONAI working group for onboarding paths, tutorials, lecture material, documentation structure, and community engagement." canonical: https://project-monai.github.io/wg_outreach_adoption_education.html audience: [engineer] last_updated: 2026-09-08 source: wg_outreach_adoption_education.html --- # Outreach, Adoption, and Education Working Group Mission Statement This working group formed by merging the former Outreach and Adoption Working Group with the former Education Working Group. It covers both sides of that merge: the learning materials that get people started with MONAI, and the community work that brings them in. Much of the material already exists, so most of the effort goes into reorganizing and maintaining it rather than writing from scratch. ## Get Involved ### Join Our Working Group We are looking for engaged members for the MONAI Outreach, Adoption, and Education Working Group. If you are interested in contributing, even if it is just participating and adding to discussions at our meetings, please fill out our participation form and get in contact with us. [Working Group Participation Form](https://forms.gle/h1aEvJK192D61zU1A) ### Meeting Information - • **Schedule:** Every other Friday - • **Time:** 9:00 am PST / 12:00 pm EST / 6:00 pm CET - • **Contact:** Reach out through the [MONAI Slack](https://join.slack.com/t/projectmonai/shared_invite/zt-3hucgm02q-i8Bn9XofDZs2UGOH4jUl4w) ## Initiatives ### Learning Materials - • A revamp of the tutorials repository into clear learning paths - • An introductory series pairing short theory with runnable notebooks - • Reusable lecture decks in executive and technical versions - • Existing documentation reorganized for clarity ### Community and Adoption - • Announcements through the website, LinkedIn, and mailing lists - • MONAI and MedTech Days, bootcamps, workshops, and MICCAI - • Success stories from people building on MONAI - • Recruiting educators and contributors into the group ## Group Leads ![Michael Zephyr](/assets/img/people/michael-zephyr.jpg) Michael Zephyr Technical Marketing Engineering Manager NVIDIA Outreach, Adoption, and Education Working Group Chair [View Profile](https://www.linkedin.com/in/michael-zephyr) ![Supriya Thathachary](/assets/img/people/supriya-thathachary.jpg) Supriya Thathachary Product Manager, Medical AI NVIDIA Outreach, Adoption, and Education Working Group Chair ![Eric Kerfoot](/assets/img/people/eric-kerfoot.png) Eric Kerfoot Software Architect in Medical Engineering King's College London Outreach, Adoption, and Education Working Group Chair ## Resources ### Learning Resources - • [MONAI Tutorials Repository](https://github.com/Project-MONAI/tutorials) - • [Tutorials and Bootcamp Materials](tutorials.html) - • [Documentation](https://monai.readthedocs.io/en/stable/) - • [MONAI on YouTube](https://www.youtube.com/c/Project-MONAI) ### Community Resources - • [MONAI Blog](https://monai.medium.com/) - • [GitHub Discussions](https://github.com/Project-MONAI/MONAI/discussions) - • [MONAI Slack](https://join.slack.com/t/projectmonai/shared_invite/zt-3hucgm02q-i8Bn9XofDZs2UGOH4jUl4w) - • [Success Stories](successstories.html) ## Collaboration Opportunities ### Content Contribution - • Write or review a tutorial notebook - • Contribute slides you already teach with - • Fix or reorganize a documentation page ### Community Engagement - • [Join our regular working group meetings](https://forms.gle/h1aEvJK192D61zU1A) - • Run a session at a workshop or bootcamp - • Tell us where MONAI is in use at your company or center --- --- title: "MONAI - Ultrasound Working Group" description: "Advancing MONAI's support for ultrasound AI research and development, from point-of-care diagnosis to surgical robot guidance and from raw RF to 4D acquisitions." canonical: https://project-monai.github.io/wg_ultrasound.html audience: [engineer] last_updated: 2026-09-08 source: wg_ultrasound.html --- # Ultrasound Working Group Mission Statement This working group will focus on enabling reproducible, scalable AI development for ultrasound within the MONAI ecosystem. Key priorities include harmonizing data formats (e.g., DICOM for B-mode); supporting data streaming (e.g., for RF signals); standardizing annotation formats and labeling protocols for common ultrasound tasks; and defining reusable pipelines for training, inference, evaluation, and deployment. Short-term and long-term priorities and focal clinical applications will be determined by participants of the group. By collaborating with the MONAI developers, the group will move from requirements to implemented solutions. ## Initiatives ### Broad, long-term initiatives - • Connectivity with ultrasound devices - • Data de-identification and pre-processing - • TotalSegmentator for ultrasound - • Algorithms for B-mode and RF ultrasound - • Ultrasound simulation ### Current Focus - • TBD ### Future Focus - • TBD ## Group Leads ![Tina Kapur](/assets/img/people/tina-kapur.jpg) Tina Kapur Center Executive Director, Surgical Planning Laboratory Department of Radiology, Brigham and Women's Hospital, Harvard Medical School Ultrasound Working Group Chair [View Profile](https://www.linkedin.com/in/tina-kapur-5a91a8139/) ![Stephen Aylward](/assets/img/people/stephen-aylward.jpg) Stephen Aylward Global Lead for Strategic Applied Research, Medical Devices NVIDIA Ultrasound Working Group Chair [View Profile](https://www.linkedin.com/in/stephenaylward/) ## Collaboration Opportunities ### Join and Contribute - • [Join our regular working group meetings](https://forms.gle/TVRKXbonDXvTuzEc7) - • Share your ultrasound experience and requirements - • Participate in defining standards and best practices - • Contribute your ultrasound AI methods - • Help shape the future of medical AI for ultrasound --- --- title: "MONAI - Success Stories" description: "How healthcare institutions and industry partners run MONAI in production. Read real-world case studies and success stories." canonical: https://project-monai.github.io/successstories.html audience: [engineer] last_updated: 2026-09-08 source: successstories.html --- # Success Stories Production deployments of MONAI across three continents: Mayo Clinic Florida integrating AI into clinical radiology, Siemens Healthineers distributing models through its Digital Marketplace, CAI2R adding MAP support to the mercure orchestrator, and NHS trusts deploying at scale with AIDE. Each story covers the integration pattern, what it took to ship, and what changed in the clinic. Case Studies ## How Institutions Use MONAI ![Mayo Clinic](/assets/img/logos/mayo-clinic.png) Featured ### Pioneering Clinical AI Integration Mayo Clinic Florida's Center for Augmented Intelligence in Imaging integrates MONAI-compatible AI models into clinical radiology workflows. Their infrastructure lets healthcare teams deploy imaging AI with minimal disruption to existing processes. Key Achievement Integrated with existing IT infrastructure [Read Case Study](mayo-case-study.html) [![MONAI models running in the CAII clinical viewer at Mayo Clinic](/assets/img/figures/mayo-case-study-figure-3-thumb.jpg) Mayo Clinic Florida · CAII viewer](mayo-case-study.html) [![Mayo Clinic](/assets/img/logos/mayo-clinic.png) × ![Siemens Healthineers](/assets/img/logos/siemens.png) Scalable Deployment ### Mayo Clinic AI, Available to 10,000+ Institutions MONAI-based tools built at one institution, distributed worldwide through the Siemens Healthineers Digital Marketplace. Researchers install them without writing code. Key Achievement 10,000+ institutions enabled with 1-hour deployment](mayo-siemens-case-study.html) [![Siemens Healthineers](/assets/img/logos/siemens.png) Enterprise Integration ### Accelerating AI Integration at Enterprise Scale MONAI Deploy integrated into Syngo Carbon and syngo.via. Healthcare providers install trained AI models from the Digital Marketplace in a few clicks. Key Achievement Deployment time cut from months to minutes](https://blogs.nvidia.com/blog/rsna-siemens-healthineers-monai-medical-imaging-ai/) [![mercure](/assets/img/logos/mercure.png) Open Source Platform ### Rapid MAP Deployment with DICOM Orchestration CAI2R added MONAI Application Package support to the mercure DICOM Orchestrator: web-based configuration, monitoring, and routing for clinical AI. Key Feature Open-source DICOM orchestration with MAP support](https://monai.medium.com/rapid-deployment-of-monai-application-packages-maps-in-radiology-workflows-using-the-mercure-fe7cfd77acce) [![AI Centre for Value Based Healthcare](/assets/img/logos/aicentre.jpg) Clinical Platform ### Deploying AI Across the NHS with AIDE The AI Centre's AIDE platform lets NHS trusts deploy AI models at scale: connected to patient records, with results delivered into clinical systems. Key Innovation Real-time AI analysis integrated with EPR systems](https://www.aicentre.co.uk/our-platforms#tab-1) Get Involved ## Share Your Success Story Are you running MONAI in production? We'd love to feature your implementation and share it with the community. [Share Your Story](https://github.com/Project-MONAI/project-monai.github.io/issues) --- --- title: "MONAI - Mayo Clinic Case Study" description: "Learn how Mayo Clinic's Center for Augmented Intelligence in Imaging (CAII) uses MONAI to integrate AI models within clinical-imaging workflows." canonical: https://project-monai.github.io/mayo-case-study.html audience: [clinician, engineer] last_updated: 2026-09-08 source: mayo-case-study.html --- ![Mayo Clinic](/assets/img/logos/mayo-clinic.png) MONAI Case Study # Center for Augmented Intelligence in Imaging Mayo Clinic Florida Integrating and Deploying AI Models within Clinical-Imaging Workflows At a glance Institution Mayo Clinic Florida Center for Augmented Intelligence in Imaging Domain Radiology: chest X-ray, mammography, brain MRI, coronary CTA MONAI components [MONAI Core](core.html) [MONAI Deploy](deploy.html) Scale 10+ AI capabilities in clinical workflows, multi-site and real-time Integration DICOM, FHIR, HL7, IHE; on-premise, cloud, and hybrid Overview ## Clinical AI Integration Effective integration of imaging-related (pixel- and nonpixel-based) Artificial Intelligence (AI) models into existing clinical Radiology workflows is critical since such additions can greatly impact (either positively or negatively) operational efficiencies or downstream decision making (e.g., surgery, pathology, interventions, and drug precautions) \[1\]. In order to integrate imaging-AI capabilities with minimal negative influence on existing Radiology workflows (Figure 1), the Center for Augmented Intelligence in Imaging (CAII) at Mayo Clinic Florida has developed infrastructure and modular software packages functionally compatible with MONAI \[2\] software packages (e.g., "MONAI Core" and "MONAI Deploy"). [![Clinical workflow diagram](/assets/img/figures/mayo-case-study-figure-1A.png)](/assets/img/figures/mayo-case-study-figure-1A.png) Figure 1: A representative workflow (modeled after IHE Scheduled Workflow) shows an examination order being generated, image data being acquired during patient scanning, produced images being evaluated by a radiologist, and a report being generated by the image interpreter and then forwarded to the referring clinician for review. Clinician reviews leading to ordered biopsies or surgical interventions, may result in associated digital pathology on excised tissue samples. Challenges ## Bridging AI and Radiology AI-based infrastructure should be both indistinguishable from the existing IT environment and require, at most, minimal training of Radiology users (e.g., radiologists and technologists). Nevertheless, the introduction of such tools requires the fostering of trust among the users as well as beneficiaries (e.g., patients and referring clinicians). As the leading discipline in utilizing AI in medicine, Radiology has already recognized the need for greater efficiencies in all aspects of imaging-AI application, including AI-model development, deployment, and adaptation to real-world encounters. Unfortunately, these processes remain prohibitively time-consuming, laborious, and costly, often resulting in significant limitations to meaningful imaging-AI use (Figure 2). [![AI project development timeline](/assets/img/figures/mayo-case-study-figure-2.png)](/assets/img/figures/mayo-case-study-figure-2.png) Figure 2: Typical AI project development and time commitments [![Example use cases](/assets/img/figures/mayo-case-study-figure-3.png)](/assets/img/figures/mayo-case-study-figure-3.png) Figure 3: Example use-cases: (a) MRI-unsafe device detection on chest x-ray, (b) Breast-density classification on mammography, (c) White matter disease segmentation on MRI, (d) Segmental coronary artery stenosis detection vs. exclusion on coronary CTA [![CAII infrastructure diagram](/assets/img/figures/mayo-case-study-figure-4.png)](/assets/img/figures/mayo-case-study-figure-4.png) Figure 4: CAII infrastructure and software packages demonstrating flexible deployment options for on-premise, cloud, and hybrid environments Infrastructure ## CAII Capabilities Engineers, imaging scientists, and physicians working in the CAII have developed infrastructure and containerized software packages that integrate imaging-AI models into the existing IT environment of a busy Department of Radiology \[3-9\]. The necessary interfaces and packages can be deployed on-premise, in-cloud, or in hybrid settings (Figure 4). The goal is to require minimal user training and IT support while fostering confidence in users and beneficiaries. CAII at Mayo Clinic Florida has developed various capabilities for integrating imaging AI models into Radiology workflows. These capabilities include: - Critical-results alerting - Expert-in-the-loop AI-model deployment - On-demand model training in clinical settings - Real-time user inference-results adjudication with feedback in clinical settings - Monitoring of user satisfaction - Data collection for FDA approvals - Continuous Learning - Federated Learning - Standards-based communication (DICOM, FHIR, HL7, IHE) between clinical systems - Standards-based data collection regarding system and model performances Citations ## References 1. 1. Gupta V, Erdal BS, Ramirez C, Floca R, Jackson L, Genereaux B, Bryson S et al. "Current State of Community-Driven Radiological AI Deployment in Medical Imaging." arXiv preprint arXiv:2212.14177 (2022). 2. 2. Cardoso J, Li W, Brown R, Ma N, Kerfoot E, Wang Y, Murrey B et al. "MONAI: An open-source framework for deep learning in healthcare." arXiv preprint arXiv:2211.02701 (2022). 3. 3. Testagrose C, Gupta V, Erdal BS, White RD, Maxwell RW, Liu X, Kahanda I, Elfayoumy S, Klostermeyer W, Demirer M. "Impact of Concatenation of Digital Craniocaudal Mammography Images on a Deep-Learning Breast-Density Classifier Using Inception-V3 and ViT." In 2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), pp. 3399-3406. IEEE, 2022. 4. 4. White RD, Demirer M, Gupta V, Sebro RA, Kusumoto FM, Erdal BS. "Pre-deployment assessment of an AI model to assist radiologists in chest X-ray detection and identification of lead-less implanted electronic devices for pre-MRI safety screening: realized implementation needs and proposed operational solutions." Journal of Medical Imaging 9, no. 5 (2022): 054504. 5. 5. Gupta V, Demirer M, Maxwell RW, White RD, Erdal BS. "A multi-reconstruction study of breast density estimation using Deep Learning." arXiv preprint arXiv:2202.08238 (2022). 6. 6. Demirer M, White RD, Gupta V, Sebro RA, Erdal BS. "Cascading neural network methodology for artificial intelligence-assisted radiographic detection and classification of lead-less implanted electronic devices within the chest." arXiv preprint arXiv:2108.11954 (2021). 7. 7. White RD, Erdal BS, Demirer M, Gupta V, Bigelow MT, Dikici E, Candemir S, Galizia MS, Carpenter JL, O'Donnell TP, Halabi AH, Prevedello LM. Artificial Intelligence to Assist in Exclusion of Coronary Atherosclerosis During CCTA Evaluation of Chest Pain in the Emergency Department: Preparing an Application for Real-world Use. J Digit Imaging. 2021 Jun;34(3):554-571. doi: 10.1007/s10278-021-00441-6. Epub 2021 Mar 31. PMID: 33791909; PMCID: PMC8329136. 8. 8. Rockenbach MABC, Buch V, Gupta V, Kotecha GK, Laur O, Erdal BS, Yang D, Xu D, Ghoshajra BB, Flores MG, Dayan I, Roth H, White RD. Automatic detection of decreased ejection fraction and left ventricular hypertrophy on 4D cardiac CTA: Use of artificial intelligence with transfer learning to facilitate multi-site operations. Intelligence-Based Medicine. 2022; 6. 9. 9. Gupta V, Taylor C, Bonnet S, Prevedello LM, Hawley J, White RD, Flores MG, Erdal BS. Deep Learning Based Automatic Detection of Adequately Positioned Mammograms. Lecture Notes in Computer Science: Domain Adaptation and Representation Transfer, and Affordable Healthcare and AI for Resource Diverse Global Health. 2021; 12968:239-250. Media ## Talks [ 2023 #### MONAI Bootcamp 2023 Mayo Clinic's MONAI implementation showcase 32:30 - Mayo Clinic case study presentation ](https://www.youtube.com/watch?v=mpVEiNW9qtw&t=1950s)[ 2021 #### MONAI Bootcamp 2021 Introduction to MONAI Deploy in clinical settings Full presentation ](https://www.youtube.com/watch?v=pS68i8ShoOk) ### Interested in Deploying Your AI Research? Learn more about how MONAI Deploy can help scale your medical imaging AI applications across institutions worldwide. [Explore MONAI Deploy SDK](https://github.com/Project-MONAI/monai-deploy-app-sdk) Keep Reading ## More success stories [![Mayo Clinic](/assets/img/logos/mayo-clinic.png) × ![Siemens Healthineers](/assets/img/logos/siemens.png) ### Global AI Marketplace The CAII applications from this story, distributed to 10,000+ institutions through the Siemens Healthineers Digital Marketplace.](mayo-siemens-case-study.html)[ ### All success stories Siemens Healthineers, the NHS AIDE platform, mercure, and more institutions running MONAI in production. Browse stories](successstories.html) --- --- title: "MONAI - Mayo Clinic & Siemens Digital Marketplace Case Study" description: "From Institutional Development to Scalable Deployment: How Mayo Clinic's AI application was made available to researchers at over 10,000 institutions via Siemens Healthineers Digital Marketplace" canonical: https://project-monai.github.io/mayo-siemens-case-study.html audience: [clinician, engineer] last_updated: 2026-09-08 source: mayo-siemens-case-study.html --- ![Mayo Clinic](/assets/img/logos/mayo-clinic.png) ![Siemens Healthineers](/assets/img/logos/siemens.png) # From Institutional Development to Scalable Deployment Center for Augmented Intelligence in Imaging (CAII), Mayo Clinic Florida Global Research Access with MONAI Deploy and the Siemens Digital Marketplace At a glance Partners Mayo Clinic Florida with Siemens Healthineers Distribution Siemens Healthineers Digital Marketplace, via syngo.via OpenApps MONAI components [MONAI Core](core.html) [MONAI Deploy](deploy.html) Scale 10,000+ institutions with access; zero-code install in about an hour Data handling All processing on-premise; no data transmitted externally Background ## Introduction At the Mayo Clinic's Center for Augmented Intelligence in Imaging (CAII), we develop AI applications to address unmet needs in imaging research. While internal development and deployment offer opportunities to validate and refine these tools, bringing them to external research environments remains a critical barrier. This case study outlines how an AI application developed at Mayo Clinic for research purposes (using MONAI Core and MONAI Deploy) was made available to researchers at over 10,000 institutions via the Siemens Healthineers Digital Marketplace. By combining open-source tooling with scalable distribution infrastructure, this approach provides a blueprint for both developers and institutions seeking to build and access imaging AI research tools. Challenge ## From Local Pipelines to Global Access Internally, our team built and validated the AI application using MONAI for model implementation, data preprocessing, and output standardization. We relied on container-based orchestration and DICOMweb compatible CAII Viewer for integration into our clinical workflows. However, this technical stack, while effective in-house, could not be assumed to operate at other sites. Our challenge was to preserve model reproducibility while eliminating deployment overhead for external users. We needed a delivery mechanism that allowed other institutions to run the tool locally, without requiring development experience, custom infrastructure, or vendor intervention. Solution ## The Siemens Healthineers Digital Marketplace To address this challenge, we collaborated with Siemens Healthineers to publish the application through their Digital Marketplace, a platform that distributes containerized research applications for use within institutional imaging environments. By wrapping the application in a Marketplace-compatible container and using the platform's built-in rule configuration and DICOM-native integration, we enabled: - Local execution at each site, preserving data governance - Zero-code installation and configuration, via GUI-based tools - DICOM-native input/output, supporting PACS integration and traceability - Reusability across research applications, regardless of modality or model specifics All application logic, preprocessing, visualization, and data export are bundled and versioned. The application is clearly labeled for research use only. For Developers ## Developer Workflow To assist developers aiming to bring their applications to the Digital Marketplace, Siemens Healthineers provides technical guidelines covering containerization, interface integration, and DICOM conformance. Developers are encouraged to align with MONAI Deploy App SDK structure and compatibility standards. While source code remains institutionally governed, the externalization process emphasizes modularity, traceability, and platform readiness. Further developer resources and templates can be found at: - [github.com/Project-MONAI/monai-deploy-app-sdk](https://github.com/Project-MONAI/monai-deploy-app-sdk) - [project-monai.github.io](https://project-monai.github.io) - [forum.siemens-healthineers.com/#/group-feeds/syngoViaFrontierDeveloperCommunity](https://forum.siemens-healthineers.com/#/group-feeds/syngoViaFrontierDeveloperCommunity?name=Research%20Frontier%20Developer%20Community) - [marketplace.teamplay.siemens-healthineers.com/app/detail/Frontier-BreastDensity](https://marketplace.teamplay.siemens-healthineers.com/app/detail/Frontier-BreastDensity) The developer workflow for externalization follows a clear path: 1. 1. Model development and validation using MONAI Core (e.g., training, evaluation, DICOM output preparation) 2. 2. Application packaging via MONAI Deploy App SDK 3. 3. Containerization and alignment with Siemens Digital Marketplace standards 4. 4. Integration of viewer and rule logic, using DICOMweb compatible or web-native tools 5. 5. Submission and publication, including metadata, version control, and usage restrictions Once published, the application becomes discoverable to registered institutions and can be deployed without code modifications or site-specific engineering. For Research Users ## User Workflow and Data Flow For research users, the process is just as short: 1. 1. Discover the application in the Siemens Healthineers Digital Marketplace 2. 2. Install via the syngo.via OpenApps interface or equivalent local platform 3. 3. Configure rules (e.g., trigger by modality or study type) 4. 4. Execute the application automatically or manually from the viewer interface 5. 5. View and export results (e.g., segmentations, reports) within their institutional PACS All processing remains on-premise. No data is transmitted externally. Application versioning and logs ensure traceability. [![Siemens Digital Marketplace interface showing application deployment workflow](/assets/img/figures/mayo-siemens-marketplace.jpg)](/assets/img/figures/mayo-siemens-marketplace.jpg) Figure: Siemens Healthineers Digital Marketplace showing the Mayo Clinic CAII Viewer demonstrating model deployment within the viewer. Results ## Outcomes and Blueprint Implications Lessons learned from deployment included the importance of early alignment with DICOM IOD standards and close coordination during UI integration. During the containerization phase, we encountered minor adjustments to viewer interactivity and study selection workflows. These were resolved in collaboration with platform engineers to ensure consistency across installations. Over 10,000 institutions have had access to this application since its publication. Most installations complete in under one hour, with no scripting or infrastructure changes required. Feedback highlights ease of configuration, integration into existing workflows, and the ability to conduct reproducible research at scale. For developers, this case provides a pathway for externalizing research tools using open-source AI frameworks and enterprise-grade delivery platforms. For research users, it offers access to validated AI applications with minimal IT overhead. Future ## Looking Ahead Ongoing efforts include evaluating the deployed tool in multicenter research collaborations. Preliminary integration into broader research networks is underway, and future versions may support federated learning feedback, provided governance conditions allow. ### Interested in Deploying Your AI Research? Learn more about how MONAI Deploy can help scale your medical imaging AI applications across institutions worldwide. [Explore MONAI Deploy SDK](https://github.com/Project-MONAI/monai-deploy-app-sdk) Keep Reading ## More success stories [![Mayo Clinic](/assets/img/logos/mayo-clinic.png) ### Clinical AI Integration at Mayo Clinic How the CAII built the applications behind this story: infrastructure, MAP containerization, and in-workflow deployment. ](mayo-case-study.html)[ ### All success stories Siemens Healthineers, the NHS AIDE platform, mercure, and more institutions running MONAI in production. Browse stories](successstories.html) --- --- title: "How to Cite MONAI" description: "Canonical BibTeX entries for MONAI papers." canonical: https://project-monai.github.io/cite.html audience: [researcher] last_updated: 2026-09-08 source: cite.html --- # How to Cite MONAI If you use MONAI in your work, please cite the framework paper: ``` @article{cardoso2022monai, title = {MONAI: An open-source framework for deep learning in healthcare}, author = {Cardoso, M. Jorge and Li, Wenqi and Brown, Richard and Ma, Nic and Kerfoot, Eric and Wang, Yiheng and Murrey, Benjamin and Myronenko, Andriy and Zhao, Can and Yang, Dong and others}, journal = {arXiv preprint arXiv:2211.02701}, year = {2022} } ``` ## MAISI ``` @inproceedings{guo2024maisi, title = {MAISI: Medical AI for Synthetic Imaging}, author = {Guo, Pengfei and others}, booktitle = {Proceedings of MICCAI}, year = {2024} } ``` ## MONAI Bundle ``` @misc{monaibundle, title = {MONAI Bundle Specification}, author = {{Project MONAI}}, year = {2023}, url = {https://github.com/Project-MONAI/MONAI/blob/dev/monai/bundle/README.md} } ``` ## MONAI Label ``` @article{diaz2022monailabel, title = {MONAI Label: A framework for AI-assisted interactive labeling of 3D medical images}, author = {Diaz-Pinto, Andres and Alle, Sachidanand and Nath, Vishwesh and Tang, Yucheng and Ihsani, Alvin and Asad, Muhammad and others}, journal = {arXiv preprint arXiv:2203.12362}, year = {2022} } ```