CT-RespGeomLib converts CT airway masks into editable geometry
This framework creates semantically labelled pipe and junction blocks from CT airway masks.
CT-RespGeomLib is a deterministic framework. It converts CT airway masks into two complementary representations.
One representation is an editable assembly of semantically labelled pipe and junction blocks. These blocks connect through canonical shared ports. The second representation is a fused lumen surface. This surface is reconstructed from tapered implicit primitives.
A lumen surface is a reconstructed geometry of the airway. A port is a shared connection between blocks. Stenosis is a narrowing of the airway. Dilation is an enlargement of the airway.
The framework was evaluated on 42 reference airway masks from ATM22 (15 cases) and AeroPath (27 cases). It produced 37 topologically accepted reconstructions. The median port-to-loop distance was mm across 2174 matched ports.
Reconstructed and reference equivalent radii showed strong agreement with . A mean bias of $-0.414$ mm arose primarily from the circular, inscribed-radius representation. Controlled pathology modelling generated 40 variants across five patient-specific geometries.
These variants included 25-75% stenosis and 50% dilation at tracheal and main-bronchus locations. Prescribed radius changes were reproduced with a mean absolute ratio error of . Additional surface preparation may be required for numerical meshing.
Key facts
- CT-RespGeomLib converts CT airway masks into editable pipe and junction blocks.
- The framework produced 37 topologically accepted reconstructions from 42 reference masks.
- The median port-to-loop distance was mm across 2174 matched ports.
- The framework enables localised modelling of airway stenosis and dilation without re-segmentation.