FWE assumes a specific tensor scaling (thresholds on MD, step size of updates, etc)
so it would probably be best to scale the tensor at the very end of the TensorReconstructionWithEigenvalueCorrectionFilter as well as input AND output of the FWE computation.

A test should be written that compares the results of TensorReconstructionWithEigenvalueCorrectionFilter with ITK Diffusion Tensor3DReconstruction Filter on a dataset without negative eigenvalues.

Another test should be written checking with Images of different reference B values, e.g. higher b-values than reference b value.

User webechr has pushed new remote branch:

bug-17983-Adapt-ReconstrToStdRep

[4a0f1a]: Merge branch 'bug-17983-Adapt-ReconstrToStdRep'

Merged commits:

2014-08-08 17:43:09 Christian Weber [17e22c]
Stndardize Reconstruction Output

2014-08-06 10:08:20 Christian Weber [03b4c1]
Documention an MiniApps in General
and Diffusion MA in detail

User webechr has pushed new remote branch:

bug-17983-Adapt-ReconstrToStdRep-2

[edf4f5]: Merge branch 'bug-17983-Adapt-ReconstrToStdRep-2'

Merged commits:

2014-08-12 12:55:43 Christian Weber [7bc26c]
rem comments and debug output

Current Status:

Test for FWE are written.
They are compared with the Matlab Reference, as well as to their current state.
Furthermore it is tested that varying reference b values do not change the filters output.

Testing for Tensor Reconstruction is still WIP