<div dir="ltr"><div><div><div>I'm using itk::GeodesicActiveContourShapePriorLevelSetImageFilter. I read the example given in (<br><a href="http://www.itk.org/Doxygen/html/Examples_2Segmentation_2GeodesicActiveContourShapePriorLevelSetImageFilter_8cxx-example.html">http://www.itk.org/Doxygen/html/Examples_2Segmentation_2GeodesicActiveContourShapePriorLevelSetImageFilter_8cxx-example.html</a>) but I don't understand the following paragraph.<br><br>// The parameters of the distribution are user-specified. Since we<br>// assumed the principal modes have already been normalized,<br>// we set the distribution to zero mean and unit variance.<br>CostFunctionType::ArrayType mean( shape->GetNumberOfShapeParameters() );<br>CostFunctionType::ArrayType stddev( shape->GetNumberOfShapeParameters() );<br>mean.Fill( 0.0 );<br>stddev.Fill( 1.0 );<br>costFunction->SetShapeParameterMeans( mean );<br>costFunction->SetShapeParameterStandardDeviations( stddev );<br><br></div>Supposing that I have NOT normalized the principal modes, how can I calculate their means? I got principal modes using itk::ImagePCAShapeModelEstimator and I know their standard deviations are sqrt(eigen values), but I don't see how can I know their means.<br><br></div><div>Any help would be very appreciated.<br></div><div><br></div>Thanks in advance,<br><br></div>Viki.<br></div>