Hi
I was wondering if there are any known factors that affect the sulci recognition. For example does accurate white or grey+CSF segmentation play important role? I don't understand why for some subjects recognition works worse than others.
Thanks
MSh
Sulci recognition
- Jean-Francois Mangin
- Posts: 337
- Joined: Mon Mar 01, 2004 10:24 am
- Location: Neurospin, CEA, France
- Contact:
Re: Sulci recognition
Grey/CSF segmentation does not play any role.
The old method based on perceptrons is based on stochastic minimization. Usually we trigger
it a couple of time and keep the best result (very long stuff of course). There is a procedure for that in brainVISA called
parallel recognition I think. This is minimizing the amount of large failure.
But I advise you to use the new recognition procedure which is much more robust for standard brains.
The old method based on perceptrons is based on stochastic minimization. Usually we trigger
it a couple of time and keep the best result (very long stuff of course). There is a procedure for that in brainVISA called
parallel recognition I think. This is minimizing the amount of large failure.
But I advise you to use the new recognition procedure which is much more robust for standard brains.
Re: Sulci recognition
Dear Jean-Francois
Thanks very much for your prompt reply.
Would you advise using "Automatic recognition" (which is default in "T1 pipeline") or "SPAM recognition" for cases where recognition is not very successful?
Thnaks
MSh
Thanks very much for your prompt reply.
Would you advise using "Automatic recognition" (which is default in "T1 pipeline") or "SPAM recognition" for cases where recognition is not very successful?
Thnaks
MSh
- riviere
- Site Admin
- Posts: 1361
- Joined: Tue Jan 06, 2004 12:21 pm
- Location: CEA NeuroSpin, Saint Aubin, France
- Contact:
Re: Sulci recognition
Hi,
Yes, the SPAM recognition processes are the newer methods. The T1 pipeline still features the older recognition both for backward compatibility of behaviour, and because the SPAM models need additional data that should be installed in BrainVisa (see the documentation).
Denis
Yes, the SPAM recognition processes are the newer methods. The T1 pipeline still features the older recognition both for backward compatibility of behaviour, and because the SPAM models need additional data that should be installed in BrainVisa (see the documentation).
Denis