TubeTK 1.4.0

This release builds valid wheels for pypi for multiple platforms.

What's Changed

• BUG: TubeTK STATIC, even if itkTubeTK SHARED by @aylward in #167
• ENH: Bump TubeTK to v1.4.0 by @aylward in #168

Full Changelog: v1.3.7...v1.4.0

v1.3.7

What's Changed

• BUG: Incorrect class specification for InverseIntensityImageFilter by @aylward in #155
• ENH: Remove the generation of vectoritkImageF4, since ITK now provides by @aylward in #156
• ENH: Replace ITKv5_CONST with const for ITKv6 compatibility by @aylward in #158
• STYLE: Make the README filename be uppercase by @jhlegarreta in #161
• ENH: Transition project configuration to pyproject.toml by @jhlegarreta in #162
• ENH: Update to support ITK >= 5.4 and MetaIO versioning by @aylward in #163
• BUG: Remove old wasm.yml build file. by @aylward in #165
• COMP: Address warnings by @aylward in #164
• ENH: Release 1.3.7 by @aylward in #166

New Contributors

• @jhlegarreta made their first contribution in #161

Full Changelog: v1.3.6...v1.3.7

TubeTK 1.3.6

Further development is being moved to a new python-only library that uses this package:
https://github.com/aylward/pytubetk

In the future, install pytubetk to have the most effective methods, APIs, and examples.

What's Changed

• ENH: Enabling aarch64 and 2014 builds by @aylward in #137
• ENH: RegisterImages can use a loaded transform by @aylward in #139
• ENH: Update MRI demos and include an MRA-only demo by @aylward in #143
• ENH: Replace incorrectly deleted MinimalPathExtractionExport.h by @aylward in #142
• BUG: ComputeTrainingMask preserve image info. TubeExtractor count. by @aylward in #151
• ENH: Update CI process to match ITK cookiecutter style, for better coverage of platforms by @aylward in #152
• ENH: Release v1.3.6 by @aylward in #154

Full Changelog: v1.3.5...v1.3.6

vtk-last

Tagging this version as the last version that contains vtk code. The maintenance of an itk module that includes vtk has become too time consuming. Instead, methods will be refactored in a future version of this repo to not use vtk, or they will be converted to python routines in a new repo that imports itk and vtk methods.

After this version, the following will be removed:

1. sliding organ registration (to be refactored, if requested / funded)
2. anisotropic registration (to be refactored, if requested / funded)
3. vessel tortuosity computation (to be refactored, if requested / funded)
4. vessel to surface (polygon) generation (to become a python method in a new repo)

TubeTK 1.3.5

Release using v5.3.0

TubeTK 1.3.4

What's Changed

• ENH: Bump ITK to v5.3.0 by @aylward in #127

Full Changelog: v1.3.3...v1.3.4

TubeTK 1.3.3

What's Changed

• Support for Python 3.10
• Significant increase in speed for PDFSegmenter and filters that depend on it (e.g., tube seed generator)
  • User can specify an "Ignore" label, and those pixels are not used in PDF estimation or classification. This allows the background (a significant portion of most medical images) to be skipped, thereby avoiding computationally expensive feature vector generation methods being called on those voxels.

Full Changelog: v1.3.2...v1.3.3

TubeTK 1.3.2

What's Changed

• Bump itk wheel for ci by @aylward in #123

Full Changelog: v1.3.1...v1.3.2

TubeTK 1.3.1

What's Changed

• ENH: Update to use ITKv53rc4.post2 and MinimalPathExtractionv1.2.2 by @aylward in #122

Full Changelog: v1.3...v1.3.1

TubeTK 1.3

Major fix to RidgeSeedFilter, PDFSegmenter, and VesselEnhancement filters via improvements to LDA and PCA basis feature computation.

Improvements include:

1. Ability to include the addition, subtraction, multiplication, and ratio of features when computing an LDA/PCA basis of a collection of features
2. Correct computation of intensity values at multiple scales as features. The prior code stopped working with updates to ITKv5that changed how GenerateData and DynamicMultiThreadedGenerateData are used.
3. Update file format for MetaIO RidgeSeed and LDA (and PCA) reading/writing. Now correctly reads/writes statistical whitening values for basis features.
4. Corrected and accelerated computation of statistical whitening values for basis features using variables computed during basis computation.
   ENH: Can specify MinLength for vessel extraction - useful for extracting initial vessels used as training examples for vesselenhancement filters.

What's Changed

• ENH: Revert to ITKrc03 and update TubeTK to v1.2.2 by @aylward in #115
• Revert "ENH: Revert to ITKrc03 and update TubeTK to v1.2.2" by @aylward in #116
• ENH: Update http to https links where valid by @tbirdso in #117
• ENH: and BUG: Fixed RidgeSeedFilter and the classes it uses by @aylward in #118
• ENH: Bump version to 1.3 by @aylward in #119

New Contributors

• @tbirdso made their first contribution in #117

Full Changelog: v1.2.1-Fix...v1.3-fix