Hi, Mark,
Thank you so much for the clear explanation!
Best wishes,
Defeng
Quoting Mark Jenkinson <[log in to unmask]>:
> Hi,
>
> The difference is all about how to calculate the replacement value of a
> previously
> zero voxel which will now become non-zero by virtue of the dilation.
>
> Consider that at such a voxel, V, the dilation kernel includes a set of
> neighbouring voxels
> of which N contain non-zero values.
> Then the options do the following calculation and put the result into
> voxel V:
> -dilM : take the N non-zero values and find their mean
> -dilD : take the mode of the N non-zero values
> (i.e. the most common value or, if there is a tie, the
> maximum)
> -dilF : take the maximum value of the N non-zero values
>
> If you are working with binary images then all of these options do the
> same job.
> If you are working with general grey-scale intensities, then the results
> will be
> different and you need to choose what you want.
>
> All the best,
> Mark
>
>
> Defeng WANG wrote:
>
> > HI, Mark,
> >
> > Another question is what is the difference between the options -dilM
> > -dilD -dilF?
> >
> > Best wishes,
> > Defeng
> > ----- Original Message ----- From: "Mark Jenkinson" <[log in to unmask]>
> > To: <[log in to unmask]>
> > Sent: Thursday, October 18, 2007 2:58 PM
> > Subject: Re: [FSL] about morphology operations
> >
> >
> >> Yes - fslmaths will do all of these.
> >> You choose an appropriate kernel (2D, 3D, cube, sphere, etc) and
> >> then select the appropriate filtering operation (including dilation and
> >> erosion). Opening and closing are done just by doing combinations
> >> of dilation and erosion.
> >>
> >> All the best,
> >> Mark
> >>
> >>
> >> On 18 Oct 2007, at 07:39, Defeng WANG wrote:
> >>
> >>> Hi, FSLers,
> >>>
> >>> Is there any command in FSL that can perform morphology
> >>> operations, such as dilation, erosion, opening, and closing?
> >>>
> >>> Thanks.
> >>> Defeng
> >>> ----- Original Message ----- From: "Kelly" <[log in to unmask]>
> >>> To: <[log in to unmask]>
> >>> Sent: Thursday, October 18, 2007 1:05 AM
> >>> Subject: [FSL] design file question
> >>>
> >>>
> >>>> Hi FSLrs:
> >>>>
> >>>> I have a question about the correct set up for a design.mat and
> >>>> design.con file.
> >>>>
> >>>> I am trying to set up a design that will compare two pre_post
> >>>> conditions; one with a drug and one without. So, I have four groups:
> >>>>
> >>>> drug_pre
> >>>> drug_post
> >>>> placebo_pre
> >>>> placebo_post
> >>>>
> >>>> Thus, I would like to compare - all within the same design so as
> >>>> to collapse variances and possibly increase power:
> >>>>
> >>>> drug pre versus drug post
> >>>> placebo pre versus placebo post
> >>>> drug pre versus placebo pre
> >>>> drug post versus placebo post
> >>>>
> >>>> What would my design files look like? I know what they should
> >>>> look like individually for a paired test and for an unpaired
> >>>> test...but this combination has thrown me for a loop. Help!
> >>>>
> >>>>
> >>>> Also for an unequal number of subjects in drug and placebo group
> >>>> do I need to reduce this so that there is an equal number of
> >>>> subjects for both or can this design be done with unequal numbers?
> >>>> Thanks so much.
> >>>>
> >>>> Kelly Brown
> >>>> UCHSC Research Assistant
> >>>> _____________________________________________________________
> >>>> Discount Online Trading - Click Now!
> >>>> http://thirdpartyoffers.juno.com/TGL2111/fc/
> >>>> Ioyw6iiejgS3jI6fJs37sf0bUKT9fZG8KRVDgPKSDyfjbIbNQKKDNT/
> >>>
>
|