Status: Currently we have data sets that are not able to generate the statistics with planar figures.
Reason: PlanarFigureMaskGenerator::GetPrincipalAxis() (Line 369) compares the result (crossmultiplication between the normal vector of the planar and the axis of the image) against 1 and only allows a differencw of mitk::eps (which could be very pedantic).
In the concrete case the result was 0.99995574500963513 instead of 1.
The question is, if this is not close enough for the PlanarFigureMaskGenerator?