How to build a model from meshes with different number of points?

[Federica Damonte]

Hello, I did the course about SSM on future learn, I have to build a SSM of femurs but the meshes have different size and number of points. How can I manage this issue? I would like to adapt the meshes with less number of points to the most precise one and then build the model. How can I do?

Thank you

[Behzad Vafaeian]

Hi,

I am answering briefly.

The different number of points (and size) does not matter; because you will/have to write a code to deform one of them (name it the reference mesh) to fit each of the other meshes (call them target meshes).

Once you've done that, you have a reference mesh and several fitted meshes to the target meshes. The fitted meshes are not exactly equal to your target meshes in terms of shapes but they will be close enough. The accuracy depends on how much time/computational resources you want to spend.

The fitted meshes have the same number of nodes/points as that of the reference mesh, and the points are in correspondence. Therefore, you can now build your SSM.

Note that the number of points of the fitted meshes and that of the target meshes are different but they closely match each other. Please see this image:

These triangular meshes, expressing two shapes, are close to each other with some precision but have different number of points.

Best regards,

On Mon, Jul 15, 2019 at 10:16 PM Federica Damonte <f.da...@student.utwente.nl> wrote:

Hello, I did the course about SSM on future learn, I have to build a SSM of femurs but the meshes have different size and number of points. How can I manage this issue? I would like to adapt the meshes with less number of points to the most precise one and then build the model. How can I do?

Thank you

--
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 post to this group, send email to scal...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/scalismo/cd1b616b-14bf-40c8-98d0-0bd08cf8f138%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Behzad@UofA]

Federica, I am going to write about SSM using Scalismo (the library; not the scalismo LAB) in my blog. It may be useful for start. I can send you the sample data I used, If you need them.

I have written about rigid alignment:

http://empty-set.me/index.php/category/ssm/

I am going to write about kernels this week; and then on non-rigid alignment.

thanks,

On Monday, July 15, 2019 at 11:13:18 PM UTC+4:30, Behzad@UofA wrote:

Hi,

I am answering briefly.

The different number of points (and size) does not matter; because you will/have to write a code to deform one of them (name it the reference mesh) to fit each of the other meshes (call them target meshes).

Once you've done that, you have a reference mesh and several fitted meshes to the target meshes. The fitted meshes are not exactly equal to your target meshes in terms of shapes but they will be close enough. The accuracy depends on how much time/computational resources you want to spend.

The fitted meshes have the same number of nodes/points as that of the reference mesh, and the points are in correspondence. Therefore, you can now build your SSM.

Note that the number of points of the fitted meshes and that of the target meshes are different but they closely match each other. Please see this image:

These triangular meshes, expressing two shapes, are close to each other with some precision but have different number of points.

Best regards,

On Mon, Jul 15, 2019 at 10:16 PM Federica Damonte <f.da...@student.utwente.nl> wrote:

Hello, I did the course about SSM on future learn, I have to build a SSM of femurs but the meshes have different size and number of points. How can I manage this issue? I would like to adapt the meshes with less number of points to the most precise one and then build the model. How can I do?

Thank you

--
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+unsubscribe@googlegroups.com.

[Federica Damonte]

Hi, I've tried the code provided in the quick start tutorial in scalismo in order to fit a target into a reference shape, but the result gives me only a part of the entire shape? which parameter do I have to adjust?

Is it the statistical model itself or the sampler?

Thank u in advance.

Kind regards

Federica

On Monday, July 29, 2019 at 7:55:58 AM UTC+2, Behzad@UofA wrote:

Federica, I am going to write about SSM using Scalismo (the library; not the scalismo LAB) in my blog. It may be useful for start. I can send you the sample data I used, If you need them.

I have written about rigid alignment:

http://empty-set.me/index.php/category/ssm/

I am going to write about kernels this week; and then on non-rigid alignment.

thanks,

On Monday, July 15, 2019 at 11:13:18 PM UTC+4:30, Behzad@UofA wrote:

Hi,

I am answering briefly.

The different number of points (and size) does not matter; because you will/have to write a code to deform one of them (name it the reference mesh) to fit each of the other meshes (call them target meshes).

Once you've done that, you have a reference mesh and several fitted meshes to the target meshes. The fitted meshes are not exactly equal to your target meshes in terms of shapes but they will be close enough. The accuracy depends on how much time/computational resources you want to spend.

The fitted meshes have the same number of nodes/points as that of the reference mesh, and the points are in correspondence. Therefore, you can now build your SSM.

Note that the number of points of the fitted meshes and that of the target meshes are different but they closely match each other. Please see this image:

These triangular meshes, expressing two shapes, are close to each other with some precision but have different number of points.

Best regards,

On Mon, Jul 15, 2019 at 10:16 PM Federica Damonte <f.da...@student.utwente.nl> wrote:

Hello, I did the course about SSM on future learn, I have to build a SSM of femurs but the meshes have different size and number of points. How can I manage this issue? I would like to adapt the meshes with less number of points to the most precise one and then build the model. How can I do?

Thank you

--
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 scal...@googlegroups.com.

[Behzad Vafaeian]

- Which "code" did you use? could you send me the link of its tutorial page please?

I think if something goes strange with a GP model, it goes back to its kernel.

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/5d8b5d02-b56c-4c46-adc6-8d844718676a%40googlegroups.com.

[Federica Damonte]

I've attached my code since

On Monday, July 29, 2019 at 3:30:54 PM UTC+2, Behzad@UofA wrote:

- Which "code" did you use? could you send me the link of its tutorial page please?

I think if something goes strange with a GP model, it goes back to its kernel.

I've attached my code since the tutorial is part of the documents provided in scalismo lab, I don't know if it's also online.

Thank you

To unsubscribe from this group and stop receiving emails from it, send an email to scal...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/scalismo/5d8b5d02-b56c-4c46-adc6-8d844718676a%40googlegroups.com.

[Federica Damonte]

https://github.com/unibas-gravis/scalismo/wiki/quickstart , I used the same kernel as they use here

[Steven A]

Do you by any chance have a similar code doing rigid registration using the ICP method using the scalismo library on a series of files listed in a folder?

[Behzad Vafaeian]

- I used this tutorial (https://unibas-gravis.github.io/scalismo-tutorial/tutorials/tutorial12.html) and it works fine. If you are using shapes that are not already rigidly aligned and hence not using a posterior GP, the kernel you are using may not capture all possible and needed deformations to fit the reference shape to any target shape. Nevertheless, it works fine If you are using the face shape provided by the tutorial.

--

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/eea450c2-e898-44aa-ab12-b5ec0e84bea8%40googlegroups.com.

[Federica Damonte]

Hi Behzad!

do you use the first type of registration or the second one?

On Tuesday, July 30, 2019 at 8:16:27 AM UTC+2, Behzad@UofA wrote:

- I used this tutorial (https://unibas-gravis.github.io/scalismo-tutorial/tutorials/tutorial12.html) and it works fine. If you are using shapes that are not already rigidly aligned and hence not using a posterior GP, the kernel you are using may not capture all possible and needed deformations to fit the reference shape to any target shape. Nevertheless, it works fine If you are using the face shape provided by the tutorial.

On Mon, Jul 29, 2019 at 6:07 PM Federica Damonte <f.da...@student.utwente.nl> wrote:

https://github.com/unibas-gravis/scalismo/wiki/quickstart , I used the same kernel as they use here

On Monday, July 15, 2019 at 7:46:08 PM UTC+2, Federica Damonte wrote:

Hello, I did the course about SSM on future learn, I have to build a SSM of femurs but the meshes have different size and number of points. How can I manage this issue? I would like to adapt the meshes with less number of points to the most precise one and then build the model. How can I do?

Thank you

--
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 scal...@googlegroups.com.

[Behzad Vafaeian]

Hi,

I use the second type I guess. I mean non-parametric non-rigid registration.

Here is my new post in my blog using that king of registration:

http://empty-set.me/index.php/2019/08/04/ssm-scalismo-3-non-rigid-registration/

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/83e7ebd2-3e4c-48f8-b2f1-1d55cd8c405e%40googlegroups.com.