TubeTK/Events/2010.11.02: Difference between revisions
From KitwarePublic
Jump to navigationJump to search
No edit summary |
|||
| Line 36: | Line 36: | ||
** Visualizations needed (Danielle) | ** Visualizations needed (Danielle) | ||
* BWH US Machine in Eternia | * BWH US Machine in Eternia | ||
= Danielle = | |||
* Successfully registered XCAT images: | |||
[[File:2010_11_02_xcatRegistration.png]] | |||
* Anisotropic gives much better motion field | |||
* Gaussian does a bit of a better job than anisotropic | |||
* Currently doing last validations = mutual overlap / Haussdorf distance between surfaces (with Valmet) | |||
* Small questions: | |||
** Max filename size in ITK? | |||
** Valmet only in Windows... | |||
* Next priorities: TubeTK dashboard, some code cleanup, register CT data, then implement improvements from Marc's suggestions + discussion with Stephen on Friday (coarse to fine, normals as matrices to handle corners) | |||
* To think about: MICCAI 2011 (submission of full papers = March 11) | |||
Revision as of 16:17, 2 November 2010
Romain
- Implementation on the ComputeTortuosity module (only Distance Metric) ---->OK!!
- With input data = Branch-truth.tre (from tubetk/Data)
- With mask covering the whole tube (Branch.n010.mha):
- Distance Metric Tortuosity for tube section 0 - 999 is 0.0524638
- Distance Metric Tortuosity for tube section 0 - 499 is 0.0656099
- With other mask (Branch.n010.sub.mha) not covering the whole scene
- Distance Metric Tortuosity for tube section 0 - 399 is 0.0811677
- With mask covering the whole tube (Branch.n010.mha):
- With input data = Branch-truth.tre (from tubetk/Data)
- Future
- Grassroots data visualization
- Conference website
- CLANG errors in metaImage
Stephen
- TubeSpatialObject
- Bug in ComputeTangentAndNormals
- Therefore, Tube/Base/Segmentation/VTree/tubeMath.h
- Pursue refactoring of itk's SpatialObject classes
- Time? Funding?
- RadiusEstimation
- Refactoring complete, bugs remain, tests being added
- Slicer integration
- Next Priority
- USC
- Funding becoming active
- Urgent need for unsupervised method
- Grassroots
- Received data
- Romain to view in Slicer and send visualization to Danielle and Stephen
- NDA signed
- Meeting next week
- UCLA/TBI
- Received data (can't find it)
- Visualizations needed (Danielle)
- BWH US Machine in Eternia
Danielle
- Successfully registered XCAT images:
- Anisotropic gives much better motion field
- Gaussian does a bit of a better job than anisotropic
- Currently doing last validations = mutual overlap / Haussdorf distance between surfaces (with Valmet)
- Small questions:
- Max filename size in ITK?
- Valmet only in Windows...
- Next priorities: TubeTK dashboard, some code cleanup, register CT data, then implement improvements from Marc's suggestions + discussion with Stephen on Friday (coarse to fine, normals as matrices to handle corners)
- To think about: MICCAI 2011 (submission of full papers = March 11)
