> My name is Flavio, MsC student at Universidade Federal Fluminense (Níterói
> - RJ). I´ve tried brain structures segmentation using SPM routines.
> Unfortunately I haven't had success. The structures don´t apper continually
> (without cessation). I think that the problem could be the high
> signal-to-noise ratio. The image is a MRI-T1 volumetric acquision. Is there
> any method that should I try? The goal is getting a better image, reducing
> signal-to-noise ratio.
The segmentation in SPM can only deal with images with a relatively high
signal-to-noise. If the intensity differences between grey and white fall
too low, then the structures can not be seen and the segmentation will
therefore not do such a great job. There are a few other segmentation
routines around (try looking on e.g. http://www.cma.mgh.harvard.edu/iatr/ ).
Maybe one of the other algorithms works better for your data.
Best regards,
-John
|