 |
 |
Take a look in the _seg8.mat file that is generated. The means, variances and (sort of) mixing proportions are saved as the "mn", "vr" and "mg" variables. When you say "voxel-wise weights", I guess you mean the latent variables for each Gaussian. These aren't written out. For suggestions about how to get hold of the variables you are interested in, I'd suggest looking at spm_preproc_write8.m Best regards, -John On 10 September 2015 at 13:30, Steenwijk, Martijn <[log in to unmask]> wrote: > Dear all, > > > > I’m using SPM12 for tissue segmentation. > > Is it possible to obtain the final parameters of the Gaussian mixture > model for each tissue type and the corresponding voxel-wise weights? > > > > I.e., CSF is modelled using 2 Gaussians, I’m looking for their parameters. > > > > Best, Martijn > > > > >