AHM 2006:Meshing

Time and Place: 6:30-7:30pm Monday (followed by drinks and dinner).

Meet up in the main AHM meeting room.

Goal

Get together to discuss various meshing and FEM topics for NA-MIC Kit. (Slides not included since much of the material is pre-publication).

Discussion Topics

• Collaboration proposal for volumetric meshing (Nicole G, Vince M, of UIowa Steve P, Alex Y) for joint biomechanics/orthopedics

• Collaboration proposal for parallel meshing (Nikos C, Andriy F, of W&M, Nicu A, Simon W, ) for brain shift simulation.

• Ongoing collaborations with SCI on bioelectric field simulation:
  • Pediatric cardiology (Matt J, John T, of Childrens Hospital Boston), Gordon K, Steve P, Dave W Jason Shepard of SCI)
  • Electrocortical Stimulation (Sandy W, Alex G, Dave W of SCI)

• General interest in mesh generation (Bill L, Will S)

Others are welcome.

Talking points from Nikos P. Chrisochoides

Here are  two issues we would like to discuss:

1) collect the requirements (help set a forum that will collect the
   requirements) to mesh generation from different medical
   applications: input description, size, time/memory constrains,
   element type, etc. Call for contributions to the "mesh generation
   wish list" from applications people.

2) explore funding opportunities for mesh generation with NIH
   either in small collaborative proposals or larger consortium.

Thank you very much,

Nikos

--

Points from the Meeting

Peter Lyster: PAR 03106 Innovations in Biomedical Computing may be able to fund this type of development, but the case will need to be made to the reviewers. Important to get Ron to say this is a key problem to enable his NCBC efforts. (See CRISP for grant probabalistic atlas natural variation in bone structure of human hand).

Will: tie closely to the biomedical driving problem, with clear research benefit where mesh generation is a critical component.

Will: There is a tet mesh generator in VTK, can do multiple regions.

Nicu: put forth datasets for testing different methods. Make it easier to develop new techniques.

General summary:

• meshing is seen as a 'solved problem' by many researchers and funding sources, yet "there's never the right meshing tool available when you need it."
• the meshing community is diverse, with several schools of thought about the best approaches and these approaches are typically embedded in non-open tools that are closely tied to particular commercial implementations.
• the (vast) majority of meshing is oriented to CAD-style meshing of analytic surfaces. Another major area is mesh refinement.
• opportunity exists to find cooperative meshing developers who will make existing codes available for open source use.
• possible link with the Simbios NCBC at Stanford
• Biomedical Meshing is a big missing piece that could have a project built around it.
  • specialized for segmented volumetric data
  • driven by important biomedical issues (like the cardiac, neuro, and orthopedic ones listed above)