[Insight-users] ICP example data

[Luis Ibanez]

Hi Bing,

These two files were graciously contributed by the ISIS Center
at Georgetown University:

http://www.isis.georgetown.edu/PortalVBVS/DesktopDefault.aspx

As far as I recall, the points were extracted from the vascular
structure of a Liver dataset.

This is run Nightly by the test "IterativeClosestPoint2Test":

http://www.cdash.org/CDash/testSummary.php?project=2&name=IterativeClosestPoint2Test&date=2009-02-09


The outcome in Zion (Debian Linux, with gcc 4.1 is)

Number of fixed Points = 486
Number of moving Points = 99
Solution = [-1.71668, -0.113132, -2.73501, 84.4617, -309.8, -425.219]



The output in Redwall (Windows Vista, with Visual Studio 9) is:

Number of fixed Points = 486
Number of moving Points = 99
Solution = [-1.73212, -0.0874141, -2.78664, 74.5081, -315.42, -431.749]


You may want to see the outcome in the machine closest to the
platform that you are using.

Are you visualizing the points somehow ?

At this point, visualization will probably be the most useful
mechanism for checking the quality of the registration.


    Please let us know,


       Thanks


          Luis

---------------
Bing Jian wrote:
> Hi,
>    In Insight/Examples/Data, there are two files:
>      IterativeClosestPointFixedPoints2.txt
>      IterativeClosestPointMovingPoints2.txt
> containing two 3D point sets for testing the ICP algorithm implemented 
> in ITK. 
> I would like to know some additional information about this dataset,
> like who originally contributed this dataset and where were these points 
> extracted from. 
> 
> The result I got from running IterativeClosestPoint2 on this dataset is
> Solution = [-1.72668, -0.0894016, -2.74065, 81.0172, -312.198, -430.148]
> The equivalent rotation matrix from the Euler angles is
> [ -0.882593306250847  -0.060593182245878   0.466216175209009
>   -0.469932886152806   0.142940536523386  -0.871051712317070
>   -0.013861395070245  -0.987874723476476  -0.154633089741434 ]
> 
> I also saved the transformed moving point set but it seems not very
> well aligned with the fixed point set.  Is there any ground truth
> transformation for this data?
> 
> Thanks,
> Bing
> 
> 
> ------------------------------------------------------------------------
> 
> _____________________________________
> 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 ITK FAQ at: http://www.itk.org/Wiki/ITK_FAQ
> 
> Follow this link to subscribe/unsubscribe:
> http://www.itk.org/mailman/listinfo/insight-users