[vtkusers] 3 speed questions

Jonathan Morra jonmorra at gmail.com
Tue Nov 9 10:51:12 EST 2010


I'm sorry, I thought that Rob was working on this speed issue, if that's not
the case then I'll see if I can bury the speed hit somewhere in the program.

On Mon, Nov 8, 2010 at 11:11 PM, Karthik Krishnan <
karthik.krishnan at kitware.com> wrote:

> Robert :
>
> I'm not sure how the locator can improve the speed of initialization.
>
> It improves the speed of closest point queries. These occur after the
> widget is being manipulated (after it has been defined by the user,
> interactively or programmatically).
>
> Please correct me if I am wrong.
>
>
> Jonathan:
>
> 3 seconds is IMHO expected for intialization. As I've mentioned
> before, the initialization of the widget, when using a Dijkstra
> interpolator involves building the adjacency graph to enable single
> source shortest path calculations. This takes a little bit of time.
>
>
> Thanks
> --
> karthik
>
> On Tue, Nov 9, 2010 at 12:06 PM, Jonathan Morra <jonmorra at gmail.com>
> wrote:
> > I just tested it and it didn't seem to increase the speed of
> initialization
> > in my case at all.  Here's what I'm doing
> > 1.  Reslice a 3D image into a 2D plane
> > 2.  Find the gradient magnitude in 2D.
> > 3.  Find the inverse gradient magnitude.
> > 4.  Set up the Dijkstra interpolator with the grad inverse as the cost
> image
> > 5.  Set the following options for the interpolator
> >
> >  interpolator.GetDijkstraImageGeodesicPath().StopWhenEndReachedOn();
> >
> >  interpolator.GetDijkstraImageGeodesicPath().RepelPathFromVerticesOn();
> >
> >  interpolator.GetDijkstraImageGeodesicPath().SetCurvatureWeight(0.15);
> >
> >  interpolator.GetDijkstraImageGeodesicPath().SetEdgeLengthWeight(0.8);
> >
>  interpolator.GetDijkstraImageGeodesicPath().SetImageWeight(1.0);
> > 6.  Create a vtkContourWidget and initialize it with the same point
> twice.
> > Everything is fine except the initialization of the contour widget, that
> > still takes 3 seconds.
> > I changed my version so now I'm getting the latest source code from the
> git
> > head, this is the correct place to get it from right?
> > On Mon, Nov 8, 2010 at 1:04 PM, Robert Maynard <
> robert.maynard at kitware.com>
> > wrote:
> >>
> >> I have just merged the updated contour widget a couple minutes ago. The
> >> changes increase the speed of initializing the widget with large
> datasets
> >> (100K+ points ), and also increases the active point location speed.
> >>
> >> On Mon, Nov 8, 2010 at 3:53 PM, Sebastien Jourdain
> >> <sebastien.jourdain at kitware.com> wrote:
> >>>
> >>> No, we don't know yet...
> >>> Rob still need to validate its fix with our internal users...
> >>>
> >>> Seb
> >>>
> >>> On Mon, Nov 8, 2010 at 1:48 PM, Jonathan Morra <jonmorra at gmail.com>
> >>> wrote:
> >>> > Thanks for this update, do you guys have any idea when this might be
> >>> > merged?
> >>> >
> >>> >
> >>> > On Mon, Nov 8, 2010 at 10:46 AM, Sebastien Jourdain
> >>> > <sebastien.jourdain at kitware.com> wrote:
> >>> >>
> >>> >> Hi Jonathan,
> >>> >>
> >>> >> Rob is currently working on vtkContourWidget to make it much faster.
> >>> >> Right now, its work is not merged to master yet but it's available
> on
> >>> >> the git stage of VTK...
> >>> >>
> >>> >> Seb
> >>> >>
> >>> >>
> >>> >> On Mon, Nov 8, 2010 at 1:26 PM, Jonathan Morra <jonmorra at gmail.com>
> >>> >> wrote:
> >>> >> > Thanks for your response.  It turns out on my machine that that
> >>> >> > initialization is too slow.  What's happening is the following
> >>> >> > 1.  The user click indicating that they want to try a contour
> >>> >> > 2.  A new vtkContourWidget is made with the dijkstra interpolator.
> >>> >> > 3.  The contour widget is initialized as I described above with
> just
> >>> >> > two
> >>> >> > points.  Adding that second point is where the slowdown occurs.
> >>> >> > As far as point 2, I am initializing a contour widget with the
> >>> >> > output of
> >>> >> > vtkCutter, but I am subsampling when I do that (I take 7% of the
> >>> >> > points), so
> >>> >> > I only have a handful of control points.  Is there a way to
> >>> >> > frontload
> >>> >> > the
> >>> >> > load time so I only have to do it once as opposed to every time a
> >>> >> > user
> >>> >> > wants
> >>> >> > to draw a contour?  That is simply too slow for my application.
> >>> >> > Also do you have any idea why rendering the meshes is slow?  Could
> >>> >> > it be
> >>> >> > related to the point order, or is is something else?
> >>> >> > On Sat, Nov 6, 2010 at 8:54 AM, Karthik Krishnan
> >>> >> > <karthik.krishnan at kitware.com> wrote:
> >>> >> >>
> >>> >> >> On Tue, Nov 2, 2010 at 6:31 AM, Jonathan Morra <
> jonmorra at gmail.com>
> >>> >> >> wrote:
> >>> >> >> > on vtkDijkstraImageContourLineInterpolator and am able to
> >>> >> >> > successfully
> >>> >> >> > create the contour widget with live wire interpolation.
> However,
> >>> >> >> > when
> >>> >> >> > I
> >>> >> >> > first initialize the contour with the following code it is very
> >>> >> >> > slow
> >>> >> >> >     // For now, we're just initializing the data with
> >>> >> >> >     // the point that was clicked
> >>> >> >> >     vtkPoints points = new vtkPoints();
> >>> >> >> >
> >>> >> >> >     // The initial data MUST be at least a "line"
> >>> >> >> >     // by giving the same point twice we are effictively
> creating
> >>> >> >> > a
> >>> >> >> > zero
> >>> >> >> >     // length line
> >>> >> >> >     points.InsertNextPoint(lastContourControlPoint);
> >>> >> >> >     points.InsertNextPoint(lastContourControlPoint);
> >>> >> >> >     vtkPolyData initialData = new vtkPolyData();
> >>> >> >> >     initialData.SetPoints(points);
> >>> >> >> >     contourWidget.Initialize(initialData, 0);
> >>> >> >> > The line that is slow is the last line.  The weird part is that
> >>> >> >> > if I
> >>> >> >> > do
> >>> >> >> > not
> >>> >> >> > use live wire, and just use the default Bezier curve
> >>> >> >> > interpolation
> >>> >> >> > the
> >>> >> >> > initialization is instant.
> >>> >> >>
> >>> >> >> Yes. There are 2 issues here.
> >>> >> >>
> >>> >> >> 1. The dijkstra interpolator is a bit slow to start with, since
> >>> >> >> during
> >>> >> >> the time of initialization, it builds the adjacency information.
> >>> >> >> But
> >>> >> >> that's not a big deal when you are drawing on an image. The very
> >>> >> >> first
> >>> >> >> line segment placement is a bit slow (~3 seconds). After that its
> >>> >> >> fast
> >>> >> >> and interactive.
> >>> >> >>
> >>> >> >> 2. The real issue, I think, here is the fact that you are
> >>> >> >> initializing
> >>> >> >> the contour with the contour with lots of control points. How
> many
> >>> >> >> of
> >>> >> >> them are there ? As I understand, you are generating these
> control
> >>> >> >> points from the output of vtkCutter ? Perhaps you want to sample
> >>> >> >> the
> >>> >> >> input polyline and then feed those sample points as the control
> >>> >> >> points.
> >>> >> >>
> >>> >> >> Thanks
> >>> >> >> --
> >>> >> >> karthik
> >>> >> >
> >>> >> >
> >>> >> > _______________________________________________
> >>> >> > Powered by www.kitware.com
> >>> >> >
> >>> >> > Visit other Kitware open-source projects at
> >>> >> > http://www.kitware.com/opensource/opensource.html
> >>> >> >
> >>> >> > Please keep messages on-topic and check the VTK FAQ at:
> >>> >> > http://www.vtk.org/Wiki/VTK_FAQ
> >>> >> >
> >>> >> > Follow this link to subscribe/unsubscribe:
> >>> >> > http://www.vtk.org/mailman/listinfo/vtkusers
> >>> >> >
> >>> >> >
> >>> >
> >>> >
> >>
> >>
> >>
> >> --
> >> Robert Maynard
> >
> >
> > _______________________________________________
> > Powered by www.kitware.com
> >
> > Visit other Kitware open-source projects at
> > http://www.kitware.com/opensource/opensource.html
> >
> > Please keep messages on-topic and check the VTK FAQ at:
> > http://www.vtk.org/Wiki/VTK_FAQ
> >
> > Follow this link to subscribe/unsubscribe:
> > http://www.vtk.org/mailman/listinfo/vtkusers
> >
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20101109/4a57b990/attachment.htm>


More information about the vtkusers mailing list