Difference between revisions of "GeometryProcessing/Spring2009/Schedule/Spectral Processing4"
Jump to navigation
Jump to search
(New page: * Brad Loos Okay, what is isometry, as in ''symmetries of a shape which are invariant up to isometry''. Is there a more technical definition than just something that is the same about a s...) |
|||
| Line 2: | Line 2: | ||
Okay, what is isometry, as in ''symmetries of a shape which are invariant up to isometry''. Is there a more technical definition than just something that is the same about a shape even if you rotate/scale/animate it? | Okay, what is isometry, as in ''symmetries of a shape which are invariant up to isometry''. Is there a more technical definition than just something that is the same about a shape even if you rotate/scale/animate it? | ||
* Manasi Datar | |||
Is the Green's function using the distance as a 'measure' of covariance in some sense ? As in, if we had measurements at each vertex, would the covariance matrix do the same job that the Green's function does, without having to compute the eigen vectors ? | |||
Secondly, is there a concept similar to Mutual Information (used to measure similarity/correspondence while aligning images), defined for meshes ? It may not work in all cases, but could help eliminate naive one-to-one matching ! | |||
Revision as of 22:33, 5 February 2009
- Brad Loos
Okay, what is isometry, as in symmetries of a shape which are invariant up to isometry. Is there a more technical definition than just something that is the same about a shape even if you rotate/scale/animate it?
- Manasi Datar
Is the Green's function using the distance as a 'measure' of covariance in some sense ? As in, if we had measurements at each vertex, would the covariance matrix do the same job that the Green's function does, without having to compute the eigen vectors ?
Secondly, is there a concept similar to Mutual Information (used to measure similarity/correspondence while aligning images), defined for meshes ? It may not work in all cases, but could help eliminate naive one-to-one matching !