External Email - Use Caution        

Firstly, thanks for your explanation, I didn't know that it would take less than hours since I used recon-all to run 202 samples, I used parallel to run 8 images at once and it took almost 13 hours, but at the end , you can be sure of the results you have. Also it gives you a variety of  aseg statistics for subcortical and cortex regions, which would be useful if you want to use them as features for machine learning frameworks, so since you don't have the same results each time recon-all would be better solution.

Le sam. 27 août 2022 à 12:47, Giulio Siracusano <siracusanogiulio@gmail.com> a écrit :
        External Email - Use Caution       

I've tested reproducibility of the results either using ASEG (therefore
via recon-all -all) and also SAMSEG (via run_samseg).

The methods are different either in terms of reproducibility of results.

Multiple runs of ASEG (recon-all -all) provide the same result but it
requires 8+ hours to be executed (each). By comparing the results of
multiple runs using the file from a single patient, such results are
exactly the same, no changes at all.

If I use SAMSEG (which is a lot faster, execution time is less than 1h
using multi-thread execution), I get overall slightly similar results
(when compared to ASEG) except for the CSF which is very different.

The problem I've found is that if I run SAMSEG on the same patient data
and evaluate the results, these results (which I would expect they
should be exactly the same), are slightly different. I've done multiple
test, whether using single-thread or multi-thread execution. Every time
I execute run_samseg the results express variability. I think this is
because multiple calls to numpy.random in the python code of samseg.
This problem should be fixed to ensure reproducibility of the data.

I hope to be clear now.

I just wanted to understand is the results from ASEG and SAMSEG (which
is again, a lot faster) could be compared in terms of reproducibility,
even if these results are not equivalent per se.


Regards

_______________________________________________
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> .
Please note that this e-mail is not secure (encrypted).  If you do not wish to continue communication over unencrypted e-mail, please notify the sender of this message immediately.  Continuing to send or respond to e-mail after receiving this message means you understand and accept this risk and wish to continue to communicate over unencrypted e-mail.