With some DICOM series, reading a single slice via of such a series via ITK will produce negative spacing values for the third dimension.

As the third dimension spacing cannot be determined correctly for a single slice, GDCM interprets a tag that is not really reliable. In some cases this leads to a reported spacing that is correct except for its sign (-0.3 instead of 0.3).

I suggest to check for this special case in InitializeByItkImage and not output the loud MITK_ERROR message. Insted the spacing should be inverted and only logged to MITK_DEBUG.

See attached patch.