Hi Matt,
i worked a lot with frame grabbers of all kinds. OpenCV did a very good job in
terms of performance. But of course, the OpenCV VideoCapture API abstracts
completely from the underlying hardware. If you want to change something
hardware-specific, you would have to use the appropriate
the issue you reported in the previous mail below.
Best
Michael
Von: Rostislav Khlebnikov [mailto:rostislav.khlebni...@kcl.ac.uk]
Gesendet: Dienstag, 2. September 2014 17:08
An: Müller, Michael
Cc: mitk-users
Betreff: Re: [mitk-users] QSortFilterProxyModel bugfix and node deletion
performance
Hi Matt,
well I am not the author of the Micro Service infrastructure, but yes: There
should be no link time dependency for the persistence lib. It should be AUTO
loaded. However, we had some trouble getting it working, this is why the
statement in Modules/IGTUI/CMakeLists.txt survived
Hey Rostislav, I can also have a look on that. I am currently inspecting your
pull request 75. If there are no further implications due to the changed code,
I will merge it asap.
Michael
Von: Rostislav Khlebnikov [mailto:rostislav.khlebni...@kcl.ac.uk]
Gesendet: Mittwoch, 27. August 2014
Hi Rotislav,
your pull request 75 (QSortFilterProxyModel) was integrated. Wasn't that easy
though since there were some changes in the DataManagerView Class. Furthermore
there were trailinig whitespaces in your commits which were rejected by our
server side hook. As you are one of our best
Hi folks,
another approach for rendering live image data is to use the OpenCV extension
and render the images in the background of the MITK render window. Of course,
you bypass the MITK rendering engine and you leave the Geometry-System of MITK.
But you will be able to render your images at a
Hi Rotislav,
my colleagues referred me again to your mails concerning persistent helper
objects and using a proxy model for the data manager. I just went through the
conversation and at least using a proxy model for filtering out nodes makes
sense to me. Unfortunately, I don't see a pull
Dear MITK-users,
we recently reworked the integration of OpenCV in MITK and this is an E-Mail to
inform you about the updates.
First of all, we updated our superbuild to use OpenCV 2.4.6.1 instead of 2.4.2.
This was necessary to support Visual Studio 2012 builds out of the box. The new
version
...@ucl.ac.uk]
Gesendet: Dienstag, 6. August 2013 20:42
An: Müller, Michael
Cc: mitk-users@lists.sourceforge.net
Betreff: Re: Too Long Path for shell
Hi Michael,
I can confirm that if I build it with a shorter path, the MITK part does
complete.
However, did you read the attachment?
The problem
Dear Matt,
can you try building it with a shortened path and report if it's working then?
I know it's an annoying issue with the path length constraints but the
superbuild process creates very long paths. This cannot be easily changed.
Michael
-Ursprüngliche Nachricht-
Von: Clarkson,
Dear Ben,
from the module documentation page:
Build Instructions
* Install Python 2.x
* Activate MITK_USE_Python in the superbuild process of MITK
* Provide CMakehttp://docs.mitk.org/nightly-qt4/namespaceCMake.html with
the correct Python paths (if not found automatically)
* Build
p.s.: I forgot to say that the Python Service class is what you actually use to
embed python in MITK. The class documentation can be found here:
http://docs.mitk.org/nightly-qt4/classmitk_1_1PythonService.html
Von: Ben Ben [mailto:ben.anot...@gmail.com]
Gesendet: Dienstag, 9. Juli 2013 17:29
Hi Felix,
thanks for opening bug [1] and contributing the patch. We had a problem with a
crash at startup using the Python module due to repeated instantiation of the
Python environment. I didn’t tried it yet with your latest patch, but please
open a new bug for it specifiying the problem and
Welcome back Ingmar ☺
At least I can say something to persistency but I don’t if this solves your
problem.
try pref-Flush() after PutBool() if the option you want to set is not saved
persistently and this causes the problem. If the Preferences Service is not
shut regularly the preferences
15:54
An: Müller, Michael
Betreff: [mitk-users] MITK Unstructured Grid Visualization
Dear Michael,
We are using MITK in our open source visualization and processing tool - GIMIAS.
It will be really king of you to give your insight over the following
functionality of MITK:
To display solid
Hi Matt,
have a look at
http://docs.mitk.org/2012.12/classmitk_1_1OpenCVToMitkImageFilter.html in the
module OpenCVVideoSupport
We had some issues with RGB images but it should do the trick
Best
Michael
-Ursprüngliche Nachricht-
Von: Clarkson, Matt [mailto:m.clark...@ucl.ac.uk]
Ah forgot to write a response,
Did you compile the 64 bit and 32 bit sources in the same build directory, i.e.
did you switch from 32 bit build to 64 bit within Visual Studio? If this is the
case, this might me the problem. 64 and 32 bit builds in the same build
directory caused problems in
Hi Matthieu,
sorry for the late reply. The NodeDescriptorManager was my crime - I will have
a look at it today and come back to you as soon as possible.
Best, Michael
Von: matthieu dumont [mailto:matthieu.dum...@usherbrooke.ca]
Gesendet: Donnerstag, 6. Dezember 2012 15:53
An:
Hi Matthieu,
I just reviewed the whole thing. The UnknownDataNodeDescriptor is a plain
member of the DescriptorManager by intent. However, I see no reason why one
should not have access to it, so i filed a new bug
(http://bugs.mitk.org/show_bug.cgi?id=14293) and pushed a fix for it in the
Hi Matt,
please forget my last mail, I mixed some memories. I just tried a clean rebuild
of the MITK master and run make package on my Ubuntu 11.10. No warnings
appeared, all so files including the OpenCV libs were correctly bundled up and
the Workbench could be started. I do not know your
Dear Miguel,
tasks that run in parallel are always application specific- a general model for
background tasking is hard to implement.
But since the MITK workbench uses Qt and Nokia gives us some great tools for
threading, background calculations using Qt Threads is not too hard:
Dear Deqiang,
There are two different errors. OpenCV does not offer RelWithDebug mode. You
should, however, be able to build OpenCV in release mode and all other
libraries in RelWithDebug. But I am afraid superbuild wont work here because
OpenCV simply doesn`t know that configuration. You have
Hi Matt,
the DataNodeAdded event is essentially an event coming from the DataStorage and
then proxified through QmitkAbstractView. I.e. directly listen to the
DataStorage (code is given in QmitkAbstractView) will make you independent of
your view visibility state.
However, a view that should
-Ursprüngliche Nachricht-
Von: Clarkson, Matt [mailto:m.clark...@ucl.ac.uk]
Gesendet: Dienstag, 23. Oktober 2012 10:41
An: Müller, Michael
Betreff: Re: AW: Application level response to NodeAdded events
Thanks Michael,
so, not in a View then. So, is the application plugin the best place
Hi Alex,
we have Python Wrapping functionality in MITK, unfortunately a bit outdated.
More information is available here:
http://www.mitk.org/Documentation/PythonWrapping
A publication in german: http://ceur-ws.org/Vol-715/bvm2011_45.pdf
I just wrote an e-mail last week about the MITK python
Hi Matt,
the DataManager itself is not an intelligent thing, we are still working on its
AI ;)
I am not an expert how Qt manages Scrollbar policies across nested widgets, but
I will have a look at it.
Regards,
Michael
-Ursprüngliche Nachricht-
Von: Clarkson, Matt
Hi Peter,
you need to insert a vtkrenderer for your text as foreground renderer. In the
Measurment plugin we do it like that:
// instantiate the objects ...
m_MeasurementInfoRenderer = vtkRenderer::New();
m_MeasurementInfoAnnotation = vtkCornerAnnotation::New();
vtkTextProperty *textProp =
Hi Miri,
no problem, questions are always welcome.
Regarding your first question: As far as I know, there is no such functionality
in MITK yet. But you can easily write a small bootstrapping program, which
starts on the command line and uses your module. Here`s a sample CMakeLists.txt:
PROJECT
Hi Miklos,
sorry I missed you message. OpenCV has some trouble with certain FFMPEG
versions (Google will give you the best links ;) ). The superbuild currently
uses OpenCV 2.3 which should be used with the latest FFMPEG version. Thus,
either you update your FFMPEG version or you turn it off
Hi Matt,
you are also working on Ubuntu Linux, dont you? It seems as if the problem is
somehow OS-specific. I filed a bug last week for this issue
(http://bugs.mitk.org/show_bug.cgi?id=9117 ) and we will investigate it today.
Best
Michael
Von: Clarkson, Matt [mailto:m.clark...@ucl.ac.uk]
Hi Miklos,
we are using the Bugzilla Inbound Email Interface
(http://www.bugzilla.org/docs/3.0/html/api/email_in.html) to modify bugs from
within a git hook. The git hook itself is a python script which uses GitPython
to easily extract information for the incoming commits from our
Hi Matt and everyone on the list,
have a look on http://www.mitk.org/wiki/BugSquashingSeminars. There you`ll find
the history of our weekly bugsquashing seminars with a lot of topics specific
to MITK and , of course, every presentation as PDF. This is a valueable
documentation source for every
An: Müller, Michael
Cc: mitk-users
Betreff: Re: [mitk-users] helper objects are not visible
Hi Michael,
I modified the code as you said. I enabled viewing nodes without data.
I found that helper object is not the feature I needed, because after
the modification some nodes appeared
-Ursprüngliche Nachricht-
Von: Miklos Espak [mailto:esp...@gmail.com]
Gesendet: Donnerstag, 28. Juli 2011 23:00
An: Müller, Michael
Cc: mitk-users
Betreff: Re: [mitk-users] helper objects are not visible
Hi,
the name could have been anything, helper node was just an example.
I am afraid I
Hi Miklos,
the code you posted is not correct- you need to set a bool property helper
object and not the name:
mitk::DataNode::Pointer helperNode = mitk::DataNode::New();
helperNode-SetName(awesome node);
helperNode-SetBoolProperty(helper object, true);
Hi Miklos, Matt,
the DataManager context Menu is extensible by using the
QmitkDataNodeDescriptorManager class. The List of QmitkDataNodeDescriptor`s is
responsible what icons are shown for specific Data Types and which Actions show
up when you open up the context menu. At the moment you add a
Hi Sebastian,
mitk::DataNode`s should not be shared between threads. If you want to send
events from threads to the main (GUI) thread, you should consider using the
great Qt Thread API:
http://doc.trolltech.com/latest/threads.html
E.g. Qt supports a blocking mechanism for signals and slots,
Hi again Sebastian,
I experienced the same behavior some time ago. As far as I can remember the
crash was caused because the RenderWindow(s) were destroyed before the
VtkLayerController destructor was called causing an invalid pointer access.
Actually, I thought I fixed it at that time
Hi Sebastian,
I have created a bug for this issue: http://bugs.mitk.org/show_bug.cgi?id=7772.
I`ll let you know when I tested this.
Thanks for the contribution
Michael
-Ursprüngliche Nachricht-
Von: sebastian ordas [mailto:sebastian.or...@gmail.com]
Gesendet: Dienstag, 12. April 2011
Hi Thomas,
currently the DataManager does not listen to changes of the selected property
in nodes. We decided this on purpose because the DataManager cannot decide
whether he should *add* the node to the current selection or if he should just
select this single node.
Instead we are using the
Hi Lodron,
you may want to use the Qt Threading capabilities to maintain a reactive GUI
during calculation. The main advantage using Qt Threads is the support of
asynchronous events. E.g. the QThread class has signals like finished() and
terminated() which makes it really easy to synchronize
-Ursprüngliche Nachricht-
Von: Yashodhan Nevatia [mailto:yashodhan.neva...@spaceapplications.com]
Gesendet: Donnerstag, 16. September 2010 14:32
An: Müller Michael
Cc: mitk-users@lists.sourceforge.net
Betreff: Re: [mitk-users] Using a vtkCamera for Augmented Reality
Hi Micheal,
I forgot to mention
Sorry, some errors in the code snippet. Here are the correct lines
myPointSetDataNode-SetBoolProperty(show points, false); // mapper will hide
the start and end point
myPointSetDataNode-SetBoolProperty(show contour, true); // more important:
mapper will draw a line between all points in the
Hi Claudia,
i just compiled and executed a small program containing your code without any
problems:
#include mitkImage.h
#include mitkPicFileReader.h
#include mitkImageDataItem.h
int main(int argc, char* argv[])
{
mitk::PicFileReader::Pointer reader = mitk::PicFileReader::New();
const char
Hi Gerald,
well, you may want to argue with my colleagues if this behavior is more
intuitive :)
I will probably make a preference option out of it and let you know. In the
meantime, if you want to change this behavior on your own, have a look into the
class QmitkDataStorageTreeModel.
This
Chintamani [mailto:keshav.chintam...@spaceapplications.com]
Gesendet: Mittwoch, 23. Juni 2010 10:55
An: Müller Michael
Betreff: Re: WG: AW: AW: WG: mitkOpenCVSupport
Hi Michael,
I am having a specific issue related to the mitkOpenCVVideoSupport module.
I have created a plugin using the bundle generator
mitkOpenCVVideoSupportUI)
Bye,
Michael
-Ursprüngliche Nachricht-
Von: Keshav Chintamani [mailto:keshav.chintam...@spaceapplications.com]
Gesendet: Dienstag, 22. Juni 2010 16:30
An: Müller Michael
Betreff: Re: WG: mitkOpenCVSupport
Hi Michael,
I looked at the videoplayer plugin and made sure the meta
Hi Matthias,
first of all, you can use any thread implementation you like, but we dont claim
every MITK class to be thread save. The DataStorage, however, should be thread
safe by now.
I suspect the piece of code you pasted should just show the whole video stream
but this is not this is not
Hi Sven,
the only thing you have to do is to set a valid path in your plugin.xml (top
level of your bundle folder). E.g. for the DataManager you will find something
like:
view
id=org.mitk.views.datamanager
name=Datamanager
category=org.mitk.views.general
Hi Tangi,
the QmitkFunctionality is a convenience class providing methods that are
frequently used by nearly all modules in a MITK application. I just gave a
short seminar on this class:
http://www.mitk.org/wiki/BugSquashingSeminars?action=AttachFiledo=viewtarget=QmitkFunctionality.pdf
However,
Hi Nabil,
just try to delete the manifest file. I had this error too and this helped me.
Michael
Von: Nabil El Ouazzani [mailto:ouana...@googlemail.com]
Gesendet: Montag, 1. März 2010 09:46
An: mitk-users@lists.sourceforge.net
Betreff: [mitk-users] mt.exe : general error c101008a
Hi all,
when
Hi,
MITK does not have any python bindings yet. You are right. We had/have plans to
realize python bindings for MITK but as we currently don`t have the personnel
capacity to do so, we announced it as a project. Unfortunately, with low
resonance so far.
So stay tuned, maybe MITK will have this
BTW: The bug id is 3065, not 2065.
Bye
-Ursprüngliche Nachricht-
Von: Müller Michael [mailto:michael.muel...@dkfz-heidelberg.de]
Gesendet: Freitag, 20. November 2009 16:35
An: 'mitk-users@lists.sourceforge.net'
Betreff: Re: [mitk-users] bug: mitk::CoreObjectFactory::GetFileExtensions
Hi Phoenix,
this issue was fixed in revision 19554. As the openCherry framework is able to
start different applications depending on which Application Plugin is loaded
at startup you must provide the id of this application plugin. In Case of the
MITK ExtApp this is
Hello Stephan,
sorry for the delay. You were right that m_UseSeriesDetails in
DataTreeNodeFactory is used uninitialized with the consequence that the
behaviour is undefined. As a solution m_UseSeriesDetails will be set to true by
default in the constructor and may be changed by using
:44
An: mitk-users@lists.sourceforge.net
Betreff: Re: [mitk-users] Errors building svn with qt4
Hi Michael,
the line is and was activated, so this will not solve the problem.
But thanks anyway,
Juliane
Quoting Müller Michael michael.muel...@dkfz-heidelberg.de:
Hi Juliane,
it seems
56 matches
Mail list logo