[Insight-users] Transforming image to world coordinates with negative z axis

[John Drescher]

Luis,
I am Sorry. I will have to get you new output. The debug output I
posted for GetOutput()->Print( std::cout ) was not the same as the
original post.

John

On Thu, Mar 12, 2009 at 4:35 PM, Luis Ibanez <luis.ibanez at kitware.com> wrote:
> Hi John,
>
> Thanks a lot for posting the output of the Print() method.
>
> It seems that the image is not being read as the last slice
> begin located at:
>
>                Z = -359.750
>
> Instead it is located at:
>
>
>      Z = -35.685 + 2.5 * 107 =  +231.81
>
> and... even if the spacing was negative,
> it could have only got to:
>
>      Z = -35.685 - 2.5 * 107 =  -303.18
>
>
>
> This raises suspicion about what you call the
>
>
>           "last DICOM slice..."
>
>
> Are you sure that it belongs to the same Series
> as the first slice ?
>
>
> Also,
> If you visualize the image that you can read with ITK,
> does it look ok ?
>
>
> BTW: As DICOM readers/viewers you may want to try:
>
>
>    A) Slicer
>       http://www.slicer.org
>
>    B) VV
>       http://www.midasjournal.org/browse/publication/288
>
>
>
> Please let us know what you find.
>
>
>    Thanks
>
>
>       Luis
>
>
> -------------------
> John Drescher wrote:
>>>
>>> You are right, this looks strange.
>>>
>>>
>>> Please do the following:
>>>
>>> Just after you load your DICOM image into an ITK image,
>>> call the Print() method in the resulting image, and post
>>> the outcome to the mailing list.
>>>
>>> Your code should look something like
>>>
>>> imageSeriesReader->Update();
>>>
>>> imageSeriesReader->GetOutput()->Print( std::cout );
>>>
>>>
>>> The output of the Print() method will help us to locate
>>> the source of the problem.
>>>
>>>
>>
>>
>> OrientedImage (01D5C650)
>>  RTTI typeinfo:   class itk::OrientedImage<short,3>
>>  Reference Count: 1
>>  Modified Time: 2757
>>  Debug: Off
>>  Observers:
>>    none
>>  Source: (01D5C508)
>>  Source output index: 0
>>  Release Data: Off
>>  Data Released: False
>>  Global Release Data: Off
>>  PipelineMTime: 30
>>  UpdateMTime: 2758
>>  LargestPossibleRegion:
>>    Dimension: 3
>>    Index: [0, 0, 0]
>>    Size: [512, 512, 107]
>>  BufferedRegion:
>>    Dimension: 3
>>    Index: [0, 0, 0]
>>    Size: [512, 512, 107]
>>  RequestedRegion:
>>    Dimension: 3
>>    Index: [0, 0, 0]
>>    Size: [512, 512, 107]
>>  Spacing: [0.626953, 0.626953, 2.5]
>>  Origin: [-174.6, -160.5, -35.685]
>>  Direction:
>> 1 0 0
>> 0 1 0
>> 0 0 1
>>
>>  IndexToPointMatrix:
>>  0.626953 0 0
>> 0 0.626953 0
>> 0 0 2.5
>>
>>  PointToIndexMatrix:
>>  1.59502 0 0
>> 0 1.59502 0
>> 0 0 0.4
>>
>>  PixelContainer:
>>    ImportImageContainer (01D5C8A0)
>>      RTTI typeinfo:   class itk::ImportImageContainer<unsigned long,short>
>>      Reference Count: 1
>>      Modified Time: 81
>>      Debug: Off
>>      Observers:
>>        none
>>      Pointer: 03410040
>>      Container manages memory: true
>>      Size: 28049408
>>      Capacity: 28049408
>>
>>
>>> ----
>>>
>>>
>>> BTW: Why do you say that the direction cosines
>>>
>>>
>>>> 1 0 0
>>>> 0 1 0
>>>> 0 0 1
>>>
>>> look wrong to you ?
>>>
>>
>> I thought it should be
>>
>> 1 0 0
>> 0 1 0
>> 0 0 -1
>>
>> so that the z axis coordinate conversion would be handled correctly.
>>
>>
>>> Given that the direction cosines in the DICOM header
>>> are
>>>
>>>
>>>> 1.000000\0.000000\0.000000\0.000000\1.000000\0.000000
>>>
>>> It seems normal to have the Direction matrix to be an
>>> identity matrix.
>>>
>>
>> Looks like I have to do some reading.
>>
>> Thanks for your help,
>> John
>>
>



-- 
John M. Drescher