2012 Winter Project Week:GeometricMetamorphosis
Home < 2012 Winter Project Week:GeometricMetamorphosis
Estimating the infiltration / recession of pathologies independent of background deformations
- Kitware: Danielle Pace, Stephen Aylward
- UNC: Marc Niethammer
- Image registration of longitudinal neuroimaging data of traumatic brain injury (TBI), with the aim of predicting long term outcome and designing treatment plans.
- The geometric metamorphosis registration algorithm jointly estimates global background deformations and local pathology changes. This allows injury changes to be analysed without the impact of background motion.
- Quantification of "infiltrative" versus "displacing" pathology changes may be a clinically-significant biomarker, and is enabled by geometric metamorphosis's separation of total deformation into background and pathology deformations.
- Develop and implement voxel-wise estimates of infiltrating versus displacing pathology growth/contraction.
- Built CALA-TK: Cross-sectional And Longitudinal Atlas toolkit
- Fixed some small bugs to get geometric metamorphosis CLI into Slicer
- Got Slicer change tracker running on my data, with help of Andriy
- Future work is integrating geometric metamorphosis into change tracker
This work will be delivered to the NA-MIC Kit as a Slicer extension.
Marc Niethammer, Gabriel L. Hart, Danielle F. Pace, Paul M. Vespa, Andrei Irimia, John D. Van Horn, and Stephen R. Aylward, Geometric Metamorphosis, MICCAI 2011.