[Insight-users] ImageFilter problems
Bill Lorensen
bill.lorensen at gmail.com
Fri Dec 23 16:41:49 EST 2011
Put a try/catch block around your code:
try
{
your code
catch (itk::ExceptionObject& e)
{
std::cerr << "Exception detected: " << e;
return EXIT_FAILURE;
}
On Fri, Dec 23, 2011 at 4:23 PM, Fleysher, Lazar
<lazar.fleysher at mountsinai.org> wrote:
> Hello Bill
>
> Yes of course I have tried your version. As I said, I will not be
> able to use your approach in the final application, but for test
> purposes I did run your program with the same segmentation violation.
> (in my final application I will need the pointer to filtered image as I will
> do more things to it)
>
> ProcessObject::UpdateOutputData and calls inside it
> cause segmentation violation whether update() is called directly or not.
>
> Thank you very much for your help
>
> Lazar
> ________________________________________
> From: Bill Lorensen [bill.lorensen at gmail.com]
> Sent: Friday, December 23, 2011 16:09
> To: Fleysher, Lazar
> Cc: Insight-users at itk.org
> Subject: Re: [Insight-users] ImageFilter problems
>
> itk does not use intermediate storage. When you say image1 = image2 it
> is just assigning pointers.
>
> I'll try compiling and running your test program if I get a chance.
>
> Did you try my test program?
>
> On Fri, Dec 23, 2011 at 3:42 PM, Fleysher, Lazar
> <lazar.fleysher at mountsinai.org> wrote:
>> Dear Bill
>>
>> You suggest to feed the output of filter directly into the writer without
>> the use of an intermediate storage. While in my end application
>> this will not work, I have tried it here and the same thing happens.
>>
>> The segmentation violation is generated on filter->update() call...
>>
>> There is something wong with ITK or libc
>>
>> What do you think?
>>
>> Lazar
>>
>> ________________________________________
>> From: Bill Lorensen [bill.lorensen at gmail.com]
>> Sent: Friday, December 23, 2011 15:01
>> To: Fleysher, Lazar
>> Cc: Insight-users at itk.org
>> Subject: Re: [Insight-users] ImageFilter problems
>>
>> Try my attached version
>>
>> On Fri, Dec 23, 2011 at 2:48 PM, Fleysher, Lazar
>> <lazar.fleysher at mountsinai.org> wrote:
>>> Hello Bill
>>>
>>> Sorry for not being clear. Attached is the example which I am trying to work out.
>>>
>>> The way to execute is :
>>>
>>> TestFilter blah.nii.gz r16slice.nii.gz
>>>
>>> (unfortunately the image file is too large to be accepted by the mailing list)
>>>
>>> The function reads the file r16slice and applies filter to it and saves it on disk into blah.nii.gz
>>>
>>> If I do not exectute shFilter->Update(); than averageimage has no pixels just before write.
>>> If I do execute shFilter->Update(); then this causes segmentation violation.
>>> If I do not execute shFilter->Update(); and do not execute averageimage = shFilter->GetOutput();
>>> then the function works as copy and blah.nii.gz is identical to the r16slice.nii.gz....
>>>
>>> Thank you very much for your help
>>>
>>> Lazar
>>> ________________________________________
>>> From: Bill Lorensen [bill.lorensen at gmail.com]
>>> Sent: Friday, December 23, 2011 12:44
>>> To: Fleysher, Lazar
>>> Cc: Insight-users at itk.org
>>> Subject: Re: [Insight-users] ImageFilter problems
>>>
>>> itk is a demand driven pipelined system. Typically you set the input
>>> to the next filter to be the output of the previous filter. Setting
>>> the input/output chain just establishes the pipeline and does not
>>> execute it. When you do an Updtae() on the last filter, the pipeline
>>> executes.
>>>
>>> averageimage=shFilter->GetOutput() is OK, but in your first email you had said:
>>> shFilter->SetInput( averageimage );
>>> shFilter->SetSValue(0.5);
>>> // shFilter->Update();
>>> averageimage = shFilter->GetOutput();
>>>
>>> This modified averageimage was was the input to shFilter.
>>>
>>>
>>> Please post a complete small, compilable example that fails. Including
>>> snippets in e-mails is too prone to error and misinterpretation.
>>>
>>> thanks,
>>>
>>> Bill
>>>
>>> On Fri, Dec 23, 2011 at 11:39 AM, Fleysher, Lazar
>>> <lazar.fleysher at mountsinai.org> wrote:
>>>> Dear Bill Lorensen
>>>>
>>>> averageimage is decleared as float 3D and then it is read from a (NIFTI) file....
>>>>
>>>> typename ImageFileReader::Pointer reader = ImageFileReader::New();
>>>> reader->SetFileName(argv[bigimage]);
>>>> reader->Update();
>>>> averageimage=reader->GetOutput();
>>>>
>>>> If I understood you correctly, I should run filter-> Update()
>>>> than to save the image do writer->SetInput(shFilter->GetOutput());
>>>> What is wrong with doing averageimage=shFilter->GetOutput()
>>>> followed by writer->SetInput(averageimage) ?
>>>>
>>>>
>>>> Regardless, shFilter->Update() generates a segmentation violation.
>>>> If I comment out the filter part, the image gets saved in to a file
>>>> and everything seem to work correctly.
>>>>
>>>> So there is something strange with the filter ... but what?
>>>>
>>>> Here is the output of gdb. Definitely there is a problem with some threads...
>>>>
>>>>
>>>> [New Thread 2122320192 (LWP 15648)]
>>>> [New Thread 46912496267840 (LWP 15512)]
>>>> [New Thread 46913547483456 (LWP 15649)]
>>>> [New Thread 46914596063552 (LWP 15650)]
>>>> [Thread 2122320192 (LWP 15648) exited]
>>>> [Thread 46914596063552 (LWP 15650) exited]
>>>> [Thread 46913547483456 (LWP 15649) exited]
>>>>
>>>> Program received signal SIGSEGV, Segmentation fault.
>>>> [Switching to Thread 46912496267840 (LWP 15512)]
>>>> 0x0000003482c076e8 in pthread_join () from /lib64/libpthread.so.0
>>>>
>>>> Here we see that the filter created 4 threads, but could not join them correctly.
>>>> Furthermore:
>>>>
>>>> (gdb) where
>>>> #0 0x0000003482c076e8 in pthread_join () from /lib64/libpthread.so.0
>>>> #1 0x00000000007dd2b5 in itk::MultiThreader::WaitForSingleMethodThread (this=0x1a11160, threadHandle=3)
>>>> at /work/home/xxx/ANTS/ITK_SOURCE/ITK/Modules/Core/Common/src/itkMultiThreaderPThreads.cxx:210
>>>> #2 0x00000000007dd9af in itk::MultiThreader::SingleMethodExecute (this=0x1a11160)
>>>> at /work/home/xxx/ANTS/ITK_SOURCE/ITK/Modules/Core/Common/src/itkMultiThreader.cxx:332
>>>> #3 0x00000000007207fb in itk::ImageSource<itk::Image<double, 3u> >::GenerateData (this=0x19cf700)
>>>> at /work/home/xxx/ANTS/ITK_SOURCE/ITK/Modules/Core/Common/include/itkImageSource.hxx:275
>>>> #4 0x00000000007d36a6 in itk::ProcessObject::UpdateOutputData (this=0x19cf700)
>>>> at /work/home/xxx/ANTS/ITK_SOURCE/ITK/Modules/Core/Common/src/itkProcessObject.cxx:1500
>>>> #5 0x00000000007e18f8 in itk::DataObject::UpdateOutputData (this=0x19c3830) at /work/home/xxx/ANTS/ITK_SOURCE/ITK/Modules/Core/Common/src/itkDataObject.cxx:434
>>>> #6 0x0000000000705f43 in itk::ImageBase<3u>::UpdateOutputData (this=0x19c3830) at /work/home/xxx/ANTS/ITK_SOURCE/ITK/Modules/Core/Common/include/itkImageBase.hxx:285
>>>> #7 0x00000000007e1816 in itk::DataObject::Update (this=0x19c3830) at /work/home/xxx/ANTS/ITK_SOURCE/ITK/Modules/Core/Common/src/itkDataObject.cxx:359
>>>> #8 0x00000000007d3bc1 in itk::ProcessObject::Update (this=0x19cf700) at /work/home/xxx/ANTS/ITK_SOURCE/ITK/Modules/Core/Common/src/itkProcessObject.cxx:1096
>>>> #9 0x000000000077a82a in itk::OptimalSharpeningImageFilter<itk::Image<float, 3u>, itk::Image<float, 3u> >::GenerateData (this=0x19ce520)
>>>> at /work/home/xxx/ANTS/ANTS_SOURCE/ANTS/Examples/../Utilities/itkOptimalSharpeningImageFilter.hxx:142
>>>> #10 0x00000000007d36a6 in itk::ProcessObject::UpdateOutputData (this=0x19ce520)
>>>> at /work/home/xxx/ANTS/ITK_SOURCE/ITK/Modules/Core/Common/src/itkProcessObject.cxx:1500
>>>> #11 0x00000000007e18f8 in itk::DataObject::UpdateOutputData (this=0x19b52b0) at /work/home/xxx/ANTS/ITK_SOURCE/ITK/Modules/Core/Common/src/itkDataObject.cxx:434
>>>> #12 0x0000000000705f43 in itk::ImageBase<3u>::UpdateOutputData (this=0x19b52b0) at /work/home/xxx/ANTS/ITK_SOURCE/ITK/Modules/Core/Common/include/itkImageBase.hxx:285
>>>> #13 0x00000000007e1816 in itk::DataObject::Update (this=0x19b52b0) at /work/home/xxx/ANTS/ITK_SOURCE/ITK/Modules/Core/Common/src/itkDataObject.cxx:359
>>>> #14 0x00000000007d3bc1 in itk::ProcessObject::Update (this=0x19ce520) at /work/home/xxx/ANTS/ITK_SOURCE/ITK/Modules/Core/Common/src/itkProcessObject.cxx:1096
>>>> #15 0x000000000077d9ad in AverageImages1<3u, 1u> (argc=10, argv=0x7fffffffd7f8) at /work/home/xxx/ANTS/ANTS_SOURCE/ANTS/Examples/AverageImages.cxx:121
>>>> #16 0x0000000000633770 in main (argc=10, argv=0x7fffffffd7f8) at /work/home/xxx/ANTS/ANTS_SOURCE/ANTS/Examples/AverageImages.cxx:308
>>>>
>>>> Does this provide some information about what is happening????
>>>>
>>>> The system I am running on is CentOS5, I guess:
>>>>
>>>> cat /etc/redhat-release
>>>> Red Hat Enterprise Linux ES release 5
>>>>
>>>> cat /etc/redhat-release.CENTOS5.orig
>>>> CentOS release 5.4 (Final)
>>>>
>>>> yum gives me this:
>>>>
>>>> yum list|grep libc
>>>> glibc.i686 2.5-42 installed
>>>> glibc.x86_64 2.5-42 installed
>>>> glibc-common.x86_64 2.5-42 installed
>>>> glibc-devel.i386 2.5-42 installed
>>>> glibc-devel.x86_64 2.5-42 installed
>>>> glibc-headers.x86_64 2.5-42 installed
>>>>
>>>>
>>>> Is this a problem with libc6?
>>>>
>>>> Thanks
>>>>
>>>> Lazar
>>>>
>>>> ________________________________________
>>>> From: Bill Lorensen [bill.lorensen at gmail.com]
>>>> Sent: Friday, December 23, 2011 08:37
>>>> To: Fleysher, Lazar
>>>> Cc: Insight-users at itk.org
>>>> Subject: Re: [Insight-users] ImageFilter problems
>>>>
>>>> For a filter to run, you must either explicitly call Update() on it,
>>>> or a filter down stream.
>>>>
>>>> Can you show the code that creates averageimage?
>>>>
>>>> Also, you should remove the statement:
>>>> averageimage = shFilter->GetOutput();
>>>>
>>>> It is already used as input to your filter. You are probably
>>>> introducing some sort of cycle in the pipeline.
>>>>
>>>> Try
>>>> writer->SetInput(shFilter->GetOutput());
>>>>
>>>>
>>>> On Fri, Dec 23, 2011 at 8:19 AM, Fleysher, Lazar
>>>> <lazar.fleysher at mountsinai.org> wrote:
>>>>>
>>>>>
>>>>> Hello insight-users at itk.org.
>>>>>
>>>>> I am new to ITK and I am experiencing a strange problem with the image filters.
>>>>> So I was wondering if someone could shed some light on this.
>>>>>
>>>>> I have this code
>>>>>
>>>>>
>>>>> typedef itk::OptimalSharpeningImageFilter<ImageType,ImageType > sharpeningFilter;
>>>>> typename sharpeningFilter::Pointer shFilter = sharpeningFilter::New();
>>>>> shFilter->SetInput( averageimage );
>>>>> shFilter->SetSValue(0.5);
>>>>> // shFilter->Update();
>>>>> averageimage = shFilter->GetOutput();
>>>>>
>>>>>
>>>>> which should filter the averageimage.
>>>>>
>>>>> Unfortunately, when I run this code the resulting averageimage contains zero pixels.
>>>>> If I uncomment the "Update()" call (which I do not know if is needed) then
>>>>> the Update() generates a segmentation violation.
>>>>>
>>>>>
>>>>> I have ITK installed from sources using git clone git://itk.org/ITK.git from
>>>>> 2 days ago. I think it is version 4.1.
>>>>>
>>>>> Also, I have noticed that if I do not call Update() and try to save the image with
>>>>>
>>>>> typename writertype::Pointer writer = writertype::New();
>>>>> writer->SetFileName("aaa");
>>>>> writer->SetInput( averageimage );
>>>>> writer->Update();
>>>>>
>>>>> writer->Update() never returns. That is, it cannot write an empty image.
>>>>>
>>>>> Am I doing something wrong? What is going on?
>>>>>
>>>>> Thank you very much for your help
>>>>>
>>>>> Lazar
>>>>> _____________________________________
>>>>> Powered by www.kitware.com
>>>>>
>>>>> Visit other Kitware open-source projects at
>>>>> http://www.kitware.com/opensource/opensource.html
>>>>>
>>>>> Kitware offers ITK Training Courses, for more information visit:
>>>>> http://www.kitware.com/products/protraining.html
>>>>>
>>>>> Please keep messages on-topic and check the ITK FAQ at:
>>>>> http://www.itk.org/Wiki/ITK_FAQ
>>>>>
>>>>> Follow this link to subscribe/unsubscribe:
>>>>> http://www.itk.org/mailman/listinfo/insight-users
>>>>
>>>>
>>>>
>>>> --
>>>> Unpaid intern in BillsBasement at noware dot com
>>>
>>>
>>>
>>> --
>>> Unpaid intern in BillsBasement at noware dot com
>>
>>
>>
>> --
>> Unpaid intern in BillsBasement at noware dot com
>
>
>
> --
> Unpaid intern in BillsBasement at noware dot com
--
Unpaid intern in BillsBasement at noware dot com
More information about the Insight-users
mailing list