On Sun, Sep 25, 2022 at 7:07 PM Douglas N. Greve <dgreve@mgh.harvard.edu> wrote:

On 9/25/2022 5:06 AM, Mikolaj Pawlak wrote:

External Email - Use Caution

On Fri, Sep 23, 2022 at 4:05 PM Douglas N. Greve <dgreve@mgh.harvard.edu> wrote:

On 9/22/2022 8:59 AM, Mikolaj Pawlak wrote:

External Email - Use Caution

Would providing better manual segmentations of those structures of interest help you in building better solutions?
I can work with the segmentation that is produced now, but I am interested in improving the results for extracerebral tissues using multimodal imaging and better manual segmentations.

For extra cerebral, you can use the CHARM SAMSEG atlas (MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://github.com/simnibs/charm-gems). SAMSEG is already modality agnostic and multimodal, so I'm not sure what you mean there.

Thank you for this link. It helps me with the skull bone segmentation. By multimodal I mean being able to extract labels for tissues visible on volumetric T2-SPACE or CUBE T2 like for example Red Nucleus. I have T1w +T2w volumetric pairs but currently cannot add red nucleus label even though I can run current version of samseg on volumetric hi-res T2w image.

The samseg authors used 20 subjects and did not find that increasing that helped all that much (at least on dice scores)

That is promising - if the list of those subjects is public and they are part of the oasis dataset as indicated by the paper there is a way to subdivide the segmentation of already existing labels

It may be possible, but won't happen soon.

I'm glad to hear that. Thanks

Can you shed some light on the steps required for creating the current atlas?

Unfortunately not. The atlas creation software was the lowest priority of the samseg authors, so it is not easy to run. I myself find it extremely challenging to create new atlases, so I can't support helping others do that. We're trying to get it more streamlined (for you and for us:)

I see. If the process can be automated like in multi-atlas label fusion it would be a great way to gain more information about both cerebral and extracerebral tissues. Is building an atlas for T2-weighted images somewhere on the priority list?

again, samseg is already multimodal and modality independent (M3I), so this should not be necessary

The problem is with the structures that are visible on non-T1w sequences, but not on T1w. In my case I would be interested in red nucleus and substantia nigra that can be segmented with multi-atlas label fusion on volumetric T2w, but not using samseg and the current version of the template. I would imagine that one can propagate labels derived from CT to T1w MRI like in CHARM SAMSEG. Doing the same thing with T2 and T2* weighted images and propagating them to T1 would enable identifying those important mid-brain structures.

thanks

Mikolaj Pawlak

thanks,

Mikolaj
thanks,
Mikolaj
On Wed, Sep 21, 2022 at 4:01 PM Douglas N. Greve <dgreve@mgh.harvard.edu> wrote:
by template I'm assuming you mean the atlas. There is not an easy way to create a new atlas or modify the existing one (though we are developing methods to do both). For the eyes, you could probably just write something to look at the column index in the eyes. If it is less than 128, then set it to left otherwise right.

On 9/21/2022 6:01 AM, Mikolaj Pawlak wrote:
External Email - Use Caution

Dear Freesurfers,
I have been using samseg for a while and like it a lot. Thank you for this tool. How do you modify the template for this tool? Is there a guideline for this?

I would like to divide Fluid_Inside_Eyes label to left and right and improve the bone information in the template. What is the procedure?

thanks,

Mikolaj Pawlak
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Mass General Brigham Compliance HelpLine at MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://www.massgeneralbrigham.org/complianceline <MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://www.massgeneralbrigham.org/complianceline> .
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Mass General Brigham Compliance HelpLine at MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://www.massgeneralbrigham.org/complianceline <MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://www.massgeneralbrigham.org/complianceline> .
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Mass General Brigham Compliance HelpLine at MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://www.massgeneralbrigham.org/complianceline <MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://www.massgeneralbrigham.org/complianceline> .