Josh -

I did generate separate volumes for each material, then I re-ran with padding turned on because it suggested it:

I've got the following message when I ran it without padding:

Problem:  Material Transitions found on boundary.
Rerun with padding

 

I assumed that this was due to the fact that my volumes are label maps and there is no "smooth" transitions between materials (or labels), unlike the frog dataset.

Here are the datasets:

 Output Volume_1.nrrd

 Output Volume_2.nrrd

 Output Volume_3.nrrd

Best,

/Ricardo

On Tue, Sep 24, 2013 at 4:33 PM, Joshua Levine <jlevine@sci.utah.edu> wrote:

Hi Ricardo,

I'm now a bit confused.  Did you generate separate volumes for each
material, or did you just run the dataset you had with padding turned
on?

Josh

On Tue, Sep 24, 2013 at 3:44 PM, Ricardo Ortiz

<ricardo.ortiz@kitware.com> wrote:
> Joshua -
>
> I see, that should be easy enough...
> I noticed (by inspecting the .ply surface only) that this does not occur
> with the frog example (that is, "air" is not meshed).
> Is it because it already has a padded region between materials and therefore
> a padded volume is not necessary?
> Is this padded region what you called "distance field" in your first email?
>
> Regards,
>
> /Ricardo
>
>
>
> On Tue, Sep 24, 2013 at 9:23 AM, Joshua Levine <jlevine@sci.utah.edu> wrote:
>>
>> Ricardo,
>>
>> This is a good feature suggestion, but I believe one that currently is not
>> supported.  The simple reason why is that for some datasets there is not a
>> clear "air" material to exclude.
>>
>> However, that said, it's fairly easy to filter the output mesh to exclude
>> tets with a particular label.  Especially if you use .node/.ele files you
>> can just extract the lines of interest and recount.  I hope this offers a
>> quick workaround.
>>
>> Thanks a lot for these emails---these are exactly the kind of features
>> we'd like to implement to simplify future users' experiences.
>>
>> Regards,
>> Josh
>>
>> On Sep 24, 2013, at 8:37 AM, Ricardo Ortiz <ricardo.ortiz@kitware.com>
>> wrote:
>>
>> Joshua -
>>
>> Thanks for the feedback.  I got it to work with your suggestion and using
>> padded volume!
>> Although it seems that with the padded volume option "air" is meshed as
>> well.  Is there a way around this?
>>
>> Thank you,
>>
>> /Ricardo
>>
>>
>> On Sun, Sep 22, 2013 at 1:10 PM, Joshua Levine <jlevine@sci.utah.edu>
>> wrote:
>>>
>>> Hi Ricardo,
>>>
>>> For your particular dataset, you'll need a separate nrrd file for each
>>> indicator function.  You've given us the label map, where each voxel
>>> is labeled by material.  What cleaver expects is a distance field per
>>> label, as in the frog example you can find on the cleaver website.
>>>
>>> Converting between the two is not super tricky, and is in fact
>>> something we've discussed supporting in the code itself.  Jonathan
>>> Bronson might in fact have this feature hidden in a development
>>> version of Cleaver.  I believe this is the currently unimplemented
>>> "loadNRRDLabelMap()" function in release 1.5.3.
>>>
>>> Regards,
>>> Josh
>>>
>>> On Fri, Sep 20, 2013 at 9:15 AM, Ricardo Ortiz
>>> <ricardo.ortiz@kitware.com> wrote:
>>> > Hello,
>>> >
>>> > We are testing cleaver to generate a mesh on the following volumetric
>>> > dataset. I get the following output:
>>> >
>>> >> Input Dimensions: 128 x 170 x 335
>>> >> Creating Mesh with Volume Size [128, 170, 335]
>>> >> ...
>>> >> Worst Angles:
>>> >> min: 180
>>> >> max: 0
>>> >> Writing settings file: man-arm-2mm-3labels-clean.info
>>> >> ...
>>> >> Cleaning up.
>>> >> Done.
>>> >
>>> >
>>> > Note the max worst angle is 0.  There is no error message nor warning
>>> > and
>>> > the output files are pretty much empty.
>>> > Is there something I am missing parameter wise?
>>> > I am using the defaults for -as and -al.
>>> >
>>> > Any help will be very much appreciated.
>>> >
>>> > Thanks,
>>> > /Ricardo
>>> >
>>> > --
>>> > Ricardo Ortiz, Ph.D.
>>> > R&D Engineer
>>> > Kitware, Inc.
>>> > 101 East Weaver St. G4
>>> > Carrboro, NC 27510
>>> > Phone: (919) 869-8870
>>> > Fax: (919) 969-6910
>>> >
>>
>>
>>
>>
>> --
>> Ricardo Ortiz, Ph.D.
>> R&D Engineer
>> Kitware, Inc.
>> 101 East Weaver St. G4
>> Carrboro, NC 27510
>> Phone: (919) 869-8870
>> Fax: (919) 969-6910
>>
>
>
>
> --
> Ricardo Ortiz, Ph.D.
> R&D Engineer
> Kitware, Inc.
> 101 East Weaver St. G4
> Carrboro, NC 27510
> Phone: (919) 869-8870
> Fax: (919) 969-6910
>

--
Ricardo Ortiz, Ph.D.
R&D Engineer
Kitware, Inc.

101 East Weaver St. G4

Carrboro, NC 27510

Phone: (919) 869-8870

Fax: (919) 969-6910