HI Petar,

Why do you have to make Seg3D in the first place?  Just fire up windows or MacOSX and you will have a working tool without the stress.

Keeping software running in Lunix is becoming a full time job—too many flavors and sub flavors.  Was this the plan all along or just an unintended consequence of the diversity within the community? Whatever, we have to live with it and try to figure it out.  The sad reality is that we have to choose between adding features that will work in Windoze/OSX or taking tons of time to get the Liinux version working again.  Funding for the project is very tight. 

Of course, if you have some great ideas for ways to make it more portable for Linux, we would welcome your contribution—it is open source, after all (-:

Thanks for your support and great ideas.

Cheers,

Rob

On Oct 4, 2017, at 06:53, Petar Petrov <pip010@gmail.com> wrote:

Hi Jess/Block,

I have tried so many times to build Seg3D and I succeeded only 1 and for short (next version was broken again).

Already forwarded my problems to the seg3D mailing list.

I am looking for some app to help me do manual segmentation, any recommendations?

The GM interface is not thin, more than 1 cm. Maybe because of CSF? Indeed CSF is either missing (correct anatomically, just happens, especially when lying inside a scanner)

Cheers,

Petar

On Tue, Oct 3, 2017 at 12:24 AM, Jess <jess@sci.utah.edu> wrote:

Changing the resolution of the segmentation requires more than changing the header.  there is a resample tool in Seg3D that will do it.  Increasing the resolution this way will also use more memory, but I’m not sure how it compares to changing the cleaver parameters.  

The RAM usage is unfortunately large, but there isn’t much that can be done to reduce it if you have thin regions.  If there are only two tissue interfaces (concentric) each layer can be meshed separately and combined in SCIRun.  

cheers,

Jess

On Oct 2, 2017, at 12:44 PM, Petar Petrov <pip010@gmail.com> wrote:

Hi Jess,

Do you mean header only or actually resize the image?

The second option I think is out of question for Cleaver2, cause it is quite mem hungry ( 32-64GB) :)

Cheers,

Petar

On Mon, Oct 2, 2017 at 7:21 PM, Jess <jess@sci.utah.edu> wrote:

Hi Petar,

It looks like you are trying to mesh a thin region of tissue. These kind of regions always cause trouble for meshing software.  You can improve the results by increasing the resolution of the background mesh.  You can try changing the scaling factor to do this, or you can resample the segmentation in Seg3D.

cheers,
Jess

> On Oct 2, 2017, at 3:02 AM, Petar Petrov <pip010@gmail.com> wrote:
>
>
> Hello all,
>
> when I generate a mesh using cleaver2 I get alot of errors of wrong tet element orientation. Those get fixed by cleaver when you pass the -j flag
>
> The mesh is generated and I can successfully load in scirun, however when using the getFieldBoundary/GetDomainBoundary the following surface mesh is produced with plenty of holes :(
>
> Any idea how I can fix that?
>

> <MyImage.jpg>

--

All the best,
Petar Petrov
http://ppetrov.net

--

All the best,
Petar Petrov
http://ppetrov.net