MONAI Label — Intelligent Medical Image Labeling

MONAI Label — Intelligent Medical Image Labeling

MONAI Label is an intelligent open-source medical image annotation tool (distinct from the MONAI training framework). It runs as a server, exposing REST APIs that OHIF Viewer, 3D Slicer, and QuPath call to get AI-generated segmentation suggestions — cutting labeling time by up to 75% through active learning.

Key features

• Active learning loop: model learns from each correction, improves suggestions over time
• Pre-built apps for radiology (CT organs, tumors), pathology (nuclei, tissue), and endoscopy
• REST API makes it embeddable into any DICOM viewer or web app
• Supports NVIDIA GPUs via CUDA; CPU fallback for inference
• Ships with 10+ pre-trained MONAI segmentation models

Quick start

pip install monailabel
# Download a pre-built app
monailabel apps --download --name radiology --output apps/
# Download sample data and start the server
monailabel start_server --app apps/radiology --studies ./studies
# Open http://localhost:8000 in OHIF or 3D Slicer

npx ai-supply add monai-label-active-learning

Curated mirror of the open-source MONAI Label (Apache-2.0). Get it from the source.