PhysioTwin4D
Build animated medical USD workflows for NVIDIA Omniverse
PhysioTwin4D is a collection of methods, workflows, tutorials, and CLI tools for creating personalized physiological digital twins from 3D medical images. Install it, clone the repository for the tutorial scripts, and work through the cards below — each tutorial runs on downloadable data and ends with the constants to change for your own. It is not validated for clinical use: PhysioTwin4D is a research and visualization toolkit, not a medical device, and must not be used for diagnosis, treatment planning, or clinical decision-making.
Version 2026.08.0
Install and Clone
Install the package, then clone the repository — the tutorial scripts do not ship in the wheel.
01Gated 4D CT to Animated USD
Segment, register and assemble a 4D CT series into an animated OpenUSD scene.
02Finetune ICON Registration
Adapt uniGradICON to your own cohort and measure what the finetuning bought you.
03Reconstruct High-Resolution 4D CT
Register every phase to one reference and reconstruct the series at its resolution.
04CT Segmentation to VTK Surfaces
Segment one CT phase and export patient anatomy as VTK PolyData surfaces.
05VTK Surfaces to Animated USD
Convert meshes into a time-sampled USD scene for Omniverse playback.
06Create a PCA Shape Model
Turn a population of meshes into a statistical shape model and its modes.
07Fit the Shape Model to a Patient
Fit the shape model to one ungated clinical scan, PCA coefficients and all.
08Propagate the Model Through 4D
Fit each case at its reference phase and carry the mesh through every phase.
09Train a PhysicsNeMo Surrogate
Train a MeshGraphNet to predict per-vertex motion from shape and phase.
10Predict Motion With the Surrogate
Replace the registration solve with one forward pass, then export to USD.
11Score the Surrogate Against the Images
Volume and surface RMSE per lobe, plus Dice per chamber, on the held-out case.
12The Whole Inference Pipeline in One Script
Go from a gated series to an animated prediction without registering a single phase.
13Breathe and Beat a Static Clinical CT
Animate one ungated breath-hold scan with both rhythms, from two networks at once.
Documentation
Explore the rest of the docs
Installation
Set up PhysioTwin4D with CUDA extras, CPU-only options, and required system tools.
Getting Started
Run your first workflow and understand the basic CT-to-USD processing path.
Tutorials & Examples
Runnable scripts covering cardiac, lung, segmentation, and USD tasks, with the inner workflow-class calls each one makes.
CLI Workflows
Use production command-line workflows for conversion, reconstruction, modeling, and USD export.
Viewing USD Files
Set up an Omniverse Kit app with RTX rendering and open the scenes the workflows produce.
Bring Your Own Data
Point the workflows at your own DICOM, NRRD, or VTK data instead of the sample datasets.
API Reference
Browse classes and modules for workflows, segmentation, registration, USD, and utilities.
Developer Docs
Understand architecture, extension points, coordinate transforms, and implementation boundaries.
Architecture
Trace the actual workflow classes and data flow from CT inputs to USD outputs.
Contributing
Follow repository conventions for code style, testing, documentation, and pull requests.
Testing
Run the fast test suite, data-gated tutorial tests, and regression checks.
Troubleshooting
Diagnose environment, data, segmentation, registration, and USD playback issues.
Tutorial Details
See Tutorials for the recommended run order, commands, datasets, per-tutorial implementation details, and the “adapt to your data” notes that close every section.