Table of Contents
- We use C++11 in DMRITool now.
- NDArray now is more efficient with C++11.
- The codes have to be built with gcc (>= 4.8), ITK (>= 4.9). Please check the building instruction.
- Some algorithms and demos on sampling scheme design are released.
- Please check the updated tutorial on sampling scheme design.
- 1-Opt: SamplingSchemeQSpace1OptEstimation, a new greedy algorithm which refines a given initialization. It can fix a known issue of IMOC when the number of samples is large. See the tutorial.
- CNLO: OptimalSamplingSingleShellCNLO for single-shell scheme design.
- CNLO: OptimalSamplingMultiShellCNLO for multiple-shell scheme design.
- jointly use IMOC + 1-Opt + CNLO for sampling scheme design:
- Mixed Integer Linear Programming (MILP) for sampling scheme design is released in matlab. See folder
- Iterative Maximum Overlap Construction (IMOC) for sampling scheme design is released in C++. See command SamplingSchemeQSpaceIMOCEstimation.
- Add a tutorial on sampling scheme design with some demos.
- Use some scripts to generate the c++ executable list, matlab function list, run demos in tutorials, update dmritool website and doxygen.
- Add VTKPolyData.py to visualize vtk files and save the visualization into png files. It can replace vtkviewer.
- The homepage of DMRITool is moving from readthedocs to github pages,
- Using gh-pages, it is possible to automatically generate some documents, for example, the list of commands.
- As required by a user, ComputeSHCoefficientsOfDWIFromSymmetricTensor is added to compute a fiber response function from a given symmetric tensor.
- Some tips using paraview is added based on the discussion
The version v0.1.1 is released. It is mainly a bug-fix release to make the building process better.
The fist release (v0.1) includes: