PhysicsNeMo AI Surrogates
MONAI Physio trains and runs PhysicsNeMo mesh-stage models: given a subject’s shape parameters and a stage (a point in the cardiac or respiratory cycle), predict a per-vertex target on the shared template mesh. When that target is a displacement, the prediction replaces a per-phase registration solve with one forward pass - see Tutorials 9 through 13 in Tutorials.
The layer follows the same has-a shape as the rest of the workflow tier: a workflow owns the data and the artifacts, and a method object owns the network.
Class |
Role |
|---|---|
Manifests, normalization, lazy datasets, checkpoints and metadata |
|
Loads a trained model and predicts raw per-point targets |
|
Interprets 3-component targets as displacements: deformed meshes, mm error statistics, rasterized deformation fields, warped images and USD |
|
Scores those predictions per structure against the acquired frames: volume difference, Dice and surface RMSE |
|
The networks to train: MeshGraphNet or fully connected |
|
The matching networks at inference time |
|
A MeshGraphNet whose loss also prices the tissue’s strain energy. It
predicts displacement, as the others do; the stress that deformation
implies comes from |
PhysicsNeMo and PyTorch Geometric are base dependencies of monai-physio, so
installing the package is enough; it requires Python >= 3.11. The imports
still happen lazily inside the methods that need them, so import
monai_physio stays lightweight.