Publication Details




Towards Distributed 3D Adaptive Anisotropic Image-to-mesh Conversion For Biomedical Applications


Kevin Garner, Fotios Drakopoulos and Nikos Chrisochoides.


Poster in NanoBioTech 2023, July, 2023




Presented is a new Image-to-Mesh (I2M) conversion software pipeline that focuses on the tessellation of complex vascular geometries required for Computational Fluid Dynamics (CFD) simulations of hemodynamics in cerebral aneurysms. The pipeline first utilizes a sequential I2M method called CBC3D, a multi-tissue isotropic tetrahedral mesh generator that constructs an adaptive Body-Centered Cubic (BCC) mesh and that subsequently deforms the mesh surface to its corresponding physical image boundary in order to improve mesh fidelity and smoothness. In order to achieve parallel anisotropic mesh adaptation and leverage the concurrency offered by emerging high-performance computing (HPC) architectures, the surface is then extracted from the CBC3D-generated mesh and is given as input to a distributed memory mesh generation method. In this distributed method, meshing functionality is separated from performance aspects by utilizing a separate entity for each - CDT3D for mesh generation and PREMA for parallel runtime support. Preliminary results show that the distributed memory method generates meshes that are qualitatively comparable to those generated by CBC3D and that fall within the operational limits of analytical CFD software. This project’s completion is expected to yield an adaptive anisotropic parallel mesh generation pipeline that will satisfy both time and quality constraints required by CFD simulations.




  [PDF]          [BibTex] 



[Return to Publication List]