Thanks for providing your implementation of global fiber tracking. When comparing your implementation with the paper by Reisert et al., MICCAI 2009, we have the following doubts:
- mitkGibbEnergyComputer.cpp, line 203ff: It appears to us that the computation of the internal energy has its sign flipped w.r.t,
to the definition by Reisert.
- mitkGibbEnergyComputer.cpp, line 146: What is the reasoning of adding m_ParticleChemicalPotential here?
- mitkGibbEnergyComputer.cpp, line 152: In our understanding of the derivation, the energy term should be computed as:
float energy = odfVal/m_ParticleWeight-2*modelVal+D2, where D2 is some constant.
- mitkEnergyComputer.cp, lines 150ff: This function does not approximate the modified Bessel function I0, although its use in
mitkGibbEnergyComputer.cpp, line 143 supposes to do so.
Thanks for your consideration. Best wishes, Frithjof