This work package provides the software components that underpin the analysis processes.
1. To provide and to develop the machine learning tools:
* To operate on literature data to infer missing data necessary for operation of the models and/or to extract rule sets (for
example on interventional outcome) to interpret the model outputs.
* To operate on the project’s own developing databases to infer missing data necessary for operation of the models and/
or to extract rule sets (for example on interventional outcome) to interpret the model outputs.
2. To provide and to develop the models and software components:
* To yield detailed anatomical representations of an individual’s mitral and aortic valve in different opening states
together with the other required heart structures such as the left ventricle and the ascending aorta from CT and 3D TEE
images or image sequences
* To extract quantitative measures based on the segmentation results, like aortic valve opening area, aortic or mitral
annulus perimeter or mitral valve regurgitant area
* To facilitate subsequent blood flow simulations by providing suitable interfaces for the definition of boundary
conditions and mesh models with sufficient resolution for subsequent volumetric meshing
3. To provide the underpinning compartmental (zero-d, lumped parameter) models that describe the distribution of flow,
pressure, oxygen saturation and other physiological parameters in the cardiovascular circulation. These are important in
their own right, but also provide boundary conditions for the localised anatomically accurate descriptions of the valve
region, which are based on the segmentation of the medical images of the individual patient.
4. To provide and to develop the software components to support sensitivity studies and the derivation of confidence
measures from multiple operations of the models.
PHILIPS will lead the workpackage and provide the segmentation tools.
PEN will provide the machine learning tools
USFD will provide the existing models of systems physiology.
ANSYS, supported by USFD, will provide the Reduced Order Modelling tools that will underpin the solution of the
complex valve flow dynamics in clinically-tractable timescales.
TUE with USFD will implement the tools for variation and sensitivity analysis.