CGAL users discussion list

["Sebastien Loriot (GeometryFactory)" <>]

Juan Carlos Lopez Alfonso wrote:
> Hi Sebastien:
>
> First of all thank you again for your answer and your comments. I have 
> done my homework but I cant see any result when I use alpha_shape. I 
> have a set of points that you can see in mi file PTV_all.off wich is a 
> triangulation with matlab. I have tried to use Alpha_Shape to obtain a 
> triangulation of my set of point where the shape of these domain of 
> points were adjusted, but without lucky. I have read the documentation, 
> but there not examples of how I can handle the result of these alpha_shape.
>
> http://dl.dropbox.com/u/2516160/MAIN.cpp

There are several problems in your code:
First, you set alpha to 0, so you will get only all the points.
Try calling find_optimal_alpha(1) and then set_alpha to the result
for example.

The second problem is that in your output functions, you consider all
simplices where as you should only consider non-exterior simplices
(and probably use the REGULARIZED version to avoid having components
of dimension less than 3).

See the Alpha_shape_3 class documentation here:
http://www.cgal.org/Manual/latest/doc_html/cgal_manual/Alpha_shapes_3_ref/Class_Alpha_shape_3.html

See here the classification of simplices:
http://www.cgal.org/Manual/latest/doc_html/cgal_manual/Alpha_shapes_3/Chapter_main.html#Subsection_42.2.3

> http://dl.dropbox.com/u/2516160/PTV_all.off

Looking at your input, I have a question, are trying to reconstruct a
shape from polygonal contours in parallel planes?

> When I run my code, I obtain a .off or vtk file with a similar 
> triamgulation of matlab, but i want to obtain a triangulation of the 
> shape of these points not of the convex hull of them.
>
> Could you see my code and give me any suggestion or way to do this?
>
> On the other hand, in all examples of the documentation they use 
> "Exact_predicates_inexact_constructions_kernel" why do you say me that 
> use "Exact_predicates_exact_constructions_kernel" instead?

Examples are written with an inexact construction kernel because when
the data is not to nasty there is usually no problem and the computation
is faster. In you case you probably have bad elements (almost flat
tetrahedra for example) and the squared circum-radii cannot be computed
using double.

Buried here:
http://www.cgal.org/Manual/latest/doc_html/cgal_manual/Alpha_shapes_3/Chapter_main.html#Section_42.4
You can read:
"the CGAL::Alpha_shape_3<Dt> requires a kernel with exact predicates and 
exact constructions"

We have something under review to address this problem with an inexact
construction kernel.


Sebastien.

> Best regards and waiting for your help
> Juan Carlos
>
> On Mon, Nov 14, 2011 at 2:57 PM, Sebastien Loriot (GeometryFactory) 
> < <mailto:>> wrote:
>
>     Juan Carlos Lopez Alfonso wrote:
>
>         Hi Sebastien:
>
>         First of all thank you for your answer and your comments.
>         In these links you can find my file of set of points and the
>         code of my example, which is simple and exactly as the example
>         "42.5.2   Building Basic Alpha Shapes for Many Points" on the
>         Alpha_Shape charapter of Cgal documentation.
>
>         http://dl.dropbox.com/u/__2516160/MAIN.cpp
>         <http://dl.dropbox.com/u/2516160/MAIN.cpp>
>         http://dl.dropbox.com/u/__2516160/Brainstem_all.txt
>         <http://dl.dropbox.com/u/2516160/Brainstem_all.txt>
>
>
>     When replacing Exact_predicates_inexact___constructions_kernel
>     by Exact_predicates_exact___constructions_kernel, the assertion is
>     gone. Explanation is here:
>     http://www.cgal.org/FAQ.html#____inexact_NT
>     <http://www.cgal.org/FAQ.html#__inexact_NT>
>     Alpha_shapes_3 require constructions so to guarantee the result to be
>     correct you need to use a kernel with exact constructions.
>
>
>
>
>
>         About my first question, I beleived that with alpha_shape was
>         obtained an approximated surface of my set of points. In my
>         problem I have a set of points and I need that my resultant
>         triangulation has the exactly same set of points, in the surface
>         and in the interior (I want to create a triangulation of a head
>         with their interior organs).
>
>
>     I don't get your point. Are you talking about missing singular vertices
>     because of the REGULARIZED mode you used in alpha_shape? If so, use
>     Alpha_shape_3::GENERAL instead.
>
>     
> http://www.cgal.org/Manual/__latest/doc_html/cgal_manual/__Alpha_shapes_3_ref/Class___Alpha_shape_3.html
>     
> <http://www.cgal.org/Manual/latest/doc_html/cgal_manual/Alpha_shapes_3_ref/Class_Alpha_shape_3.html>
>
>     Sebastien.
>
>
>         Best regards and waiting for your help
>         Juan carlos
>
>         On Mon, Nov 14, 2011 at 2:24 PM, Sebastien Loriot
>         (GeometryFactory) 
> <
>         
> <mailto:>
>  
> <mailto:
>         
> <mailto:>>>
>  wrote:
>
>            Juan Carlos Lopez Alfonso wrote:
>
>                Hi all,
>
>                There are any form to use 3D Alpha_Shapes with a general
>                triangulation like Triangulation_3? because Delaunay
>                triangulation or Regular triangulation move my points in
>         order
>                to obtain a linear approximation of the original shape and I
>                need to have a triangulation from a fix set of points.
>
>
>            Nor Delaunay, nor Regular triangulation moves your points.
>         It's building
>            a triangulation on top of your point set, so what do you mean?
>
>
>
>                On the other hand, I have created a Delaunay Triangulation
>                (Delaunay  dt;) from a set of points and when I try to
>         create an
>                alpha_shape (Alpha_shape_3 as(dt);) I have the following
>         error:
>                 CGAL::Assertion_exception !!!
>
>
>            Did you read this?
>            http://www.cgal.org/FAQ.html#____inexact_NT
>         <http://www.cgal.org/FAQ.html#__inexact_NT>
>
>            <http://www.cgal.org/FAQ.html#__inexact_NT
>         <http://www.cgal.org/FAQ.html#inexact_NT>>
>
>            If yes, showing some code might help giving you an answer.
>
>            Sebastien.
>
>
>
>                Best regrads and thank in advance
>                Juan Carlos
>
>
>
>            --     You are currently subscribed to cgal-discuss.
>            To unsubscribe or access the archives, go to
>            https://lists-sop.inria.fr/____wws/info/cgal-discuss
>         <https://lists-sop.inria.fr/__wws/info/cgal-discuss>
>            <https://lists-sop.inria.fr/__wws/info/cgal-discuss
>         <https://lists-sop.inria.fr/wws/info/cgal-discuss>>
>
>
>
>
>     -- 
>     You are currently subscribed to cgal-discuss.
>     To unsubscribe or access the archives, go to
>     https://lists-sop.inria.fr/__wws/info/cgal-discuss
>     <https://lists-sop.inria.fr/wws/info/cgal-discuss>