Subject: CGAL users discussion list
List archive
- From: "Fastl, Thomas" <>
- To: "" <>
- Cc: Jane Tournois <>, Laurent Rineau <>, Andreas Fabri <>
- Subject: Re: [cgal-discuss] Surface Smoothing after Mesh Generation
- Date: Thu, 19 May 2016 13:46:28 +0000
- Accept-language: en-US, de-AT
- Authentication-results: mail2-smtp-roc.national.inria.fr; spf=None ; spf=Pass ; spf=Pass
- Ironport-phdr: 9a23:I/2CGB9CHU6ejv9uRHKM819IXTAuvvDOBiVQ1KB92ukcTK2v8tzYMVDF4r011RmSDdSdsKgP1beempujcFJDyK7JiGoFfp1IWk1NouQttCtkPvS4D1bmJuXhdS0wEZcKflZk+3amLRodQ56mNBXsq3G/pQQfBg/4fVIsYL+lS8iJ1Y/piqibwN76XUZhvHKFe7R8LRG7/036l/I9ps9cEJs30QbDuXBSeu5blitCLFOXmAvgtI/rpMYwuwwZgf8q9tZBXKPmZOx4COUAVHV1e1wysYfguhDHCAeO/XABSX4+kxxSAgGD4guwFsP6vSL+8+Z8wyKHJtbeTLYuWD3k4b09GzHyjyJSfRAw633TzoRMjKVeqQnr70hjw4nRaZrTOvNlYqf1ftgeA2NKGNtSAX8SSrigZpcCWrJSdd1TqJPw8h5X9UOz
- Spamdiagnosticmetadata: NSPM
- Spamdiagnosticoutput: 1:23
Hello Andreas,
the major problem is to find a balance between a reasonable amount of
smoothing (increased facet_distance) and a mesh without holes (decreased
facet_distance). I'm concerned about the steps and the kind of wrinkles on
the mesh surface, while the jagged edges don't matter (although it'd be great
if they could be reduced).
Leaving out the Lloyd and ODT smoothing acutally results in a much worse mesh
(also in terms of smoothness), so this didn't do the trick as you already
suggested. I'd really appreciate any further input on this, thanks.
Best, Thomas
________________________________________
From: Andreas Fabri
<>
Sent: Thursday, May 19, 2016 1:12:00 PM
To: Fastl, Thomas
Cc: Jane Tournois; Laurent Rineau
Subject: Re: [cgal-discuss] Surface Smoothing after Mesh Generation
Hello,
Are the "steps" one can see what you want to smooth away?
I guess you get them because your facet_distance is so low.
Is that because you only have two voxel thickness?
Or is it the zig-zag line where three tissues meet?
For that we do not have a solution yet, but plan to
work on.
Lloyd and ODT are not so much about smoothing but
about improving the quality of tetrahedra.
Concerning data, there is no chance to get something
anonymized, or just a part of the image? Under an NDA
just to get a better feeling for a typical use case?
I ask because we currently improve the mesh generation
from image data for one customer, that is it is something
we work on.
best,
andreas
On 19/05/2016 12:48, Fastl, Thomas wrote:
> Hey Andreas,
>
> Unfortunately not since we're dealing with patient data which I can't
> distribute, sorry. But please find attached a zoom of the mesh I
> generated using the parameters provided in my previous email to show
> you what I've been talking about. In the meanwhile I'll experiment
> with switching on/off the Lloyd and ODT filters and explore their
> impact on the mesh smoothness. Has anybody experience with changing
> the parameters of each of them? Thanks!
>
> Best, Thomas
>
> ________________________________________ From: Andreas Fabri
> <>
> Sent: Thursday, May 19, 2016
> 11:19:20 AM To: Fastl, Thomas Subject: Re: [cgal-discuss] Surface
> Smoothing after Mesh Generation
>
> Hi Thomas,
>
> can you provide us an example data set.
>
> best,
>
> andreas
>
> On 19/05/2016 12:16, Fastl, Thomas wrote:
>> Hey Andreas,
>>
>> we're using isotropic voxels with 100um edge length (we could
>> adapt that) and the mesh parameters used are:
>>
>> facet_angle: 30 facet_size: 0.1 facet_distance: 0.1
>> cell_radius_edge_ratio: 1.5 cell_size: 0.1
>>
>> We're dealing with quite thin structures (around 2 voxels
>> thickness) and need a high-resolution mesh. Thanks!
>>
>> Best, Thomas
>> ------------------------------------------------------------------------
>>
>>
*From:*
<>
on
>> behalf of Andreas fabri
>> <>
>> *Sent:*
>> Thursday, May 19, 2016 6:14:57 AM *To:*
>>
>> *Subject:* Re: [cgal-discuss] Surface Smoothing after Mesh
>> Generation Hello,
>>
>> What is the voxel size and what are the quality parameters you gave
>> rhe algorithm?
>>
>> Andreas
>>
>> On 18 mai 2016, at 23:57, Fastl, Thomas
>> <
>> <mailto:>>
>> wrote:
>>
>>> Hello CGAL-Users,
>>>
>>>
>>> since I'm pretty new to CGAL this question might be quite basic,
>>> but I couldn't solve this problem myself and wanted to ask for
>>> help. I've managed to generate a finite element mesh from medical
>>> imaging data using the '.inr' format and I'm super happy with the
>>> quality of the tetrahedral elements. However, the externalsurface
>>> of the mesh is quite rough and I wanted to know if there is an
>>> option to smooth it? I'm aware that this might reduce the quality
>>> of the entire mesh slightly, but I'd accept that for a more
>>> smooth surface. Currently I'm using the following code to
>>> generate the mesh:
>>>
>>>
>>> C3t3 c3t3 = CGAL::make_mesh_3<C3t3>(domain, criteria,
>>> CGAL::parameters::lloyd(), CGAL::parameters::odt(),
>>>
>>> CGAL::parameters::perturb(), CGAL::parameters::exude());
>>>
>>> Does anybody know a way to perform a smoothing step after the
>>> mesh generation in CGAL? Is this actually possible within CGAL or
>>> is the scope of the software to obtain better element quality
>>> while accepting more rough surfaces? Thank you very much in
>>> advance!
>>>
>>>
>>> Best, Thomas Fastl
>>>
>>>
>
> -- Andreas Fabri, PhD Chief Officer, GeometryFactory Editor, The CGAL
> Project
>
> phone: +33.492.954.912 skype: andreas.fabri
>
--
Andreas Fabri, PhD
Chief Officer, GeometryFactory
Editor, The CGAL Project
phone: +33.492.954.912 skype: andreas.fabri
- [cgal-discuss] Surface Smoothing after Mesh Generation, Fastl, Thomas, 05/18/2016
- Re: [cgal-discuss] Surface Smoothing after Mesh Generation, Andreas fabri, 05/19/2016
- Re: [cgal-discuss] Surface Smoothing after Mesh Generation, Fastl, Thomas, 05/19/2016
- Message not available
- Message not available
- Message not available
- Re: [cgal-discuss] Surface Smoothing after Mesh Generation, Fastl, Thomas, 05/19/2016
- Message not available
- Message not available
- Message not available
- Re: [cgal-discuss] Surface Smoothing after Mesh Generation, Fastl, Thomas, 05/19/2016
- Re: [cgal-discuss] Surface Smoothing after Mesh Generation, Andreas fabri, 05/19/2016
Archive powered by MHonArc 2.6.18.