Confusions Regarding Establishing Correspondence

27 views
Skip to first unread message

Debbie Lam

unread,
Mar 23, 2022, 4:22:13 AM3/23/22
to scalismo
Hi,

Currently, I have a PDM and 4 data meshes, which I aim to make into a statistical shape model. However, arriving at the point of establishing correspondence, I have a confusion with the operation steps:

1. Establish correspondences between reference mesh (mean of PDM) and data mesh 1
2. Establish correspondences between the new mean of PDM and data mesh 2
3. Establish correspondences between the new mean of PDM and data mesh 3
4. Establish correspondences between the new mean of PDM and data mesh 4

May I know are the steps correct?

Actually, what is the output of establishing correspondences? How do I check if correspondence has been established?

Thank you in advance!


Marcel Luethi

unread,
Mar 23, 2022, 4:31:24 AM3/23/22
to Debbie Lam, scalismo
Hi Debbie

I assume the PDM you are referring to is a model that you build using a reference mesh and an analytically defined Gaussian process (as discussed in week 4 of FutureLearn).
If this is the case then the steps are as follows:
1. Fit the PDM to data mesh 1
2. Fit the PDM to data mesh 2
3. Fit the PDM to data mesh 3
...

The fitting result in each step is the reference mesh deformed in such a way that it approximates closely the shape of the mesh to which you fit it to. You can then use these deformed meshes to build a PDM. As these approximate the original meshes, you have captured all the relevant variability in the data. As they are just deformed versions of the reference mesh, you automatically have correspondence.

Best regards,

Marcel


--
You received this message because you are subscribed to the Google Groups "scalismo" group.
To unsubscribe from this group and stop receiving emails from it, send an email to scalismo+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/scalismo/16edf4d0-e0c7-49f7-9ad8-e00934821e15n%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages