Skip to content

Elaeomyrmex

Compare
Choose a tag to compare
@cookpa cookpa released this 06 Jul 14:40
· 336 commits to master since this release

Highlights (see full list of PRs below)

Bug fixes:

  • Fix Docker container bug affecting antsAtroposN4 convergence (#1334).
  • Fix ImageMath function ExtractComponentFrom3DTensor.
  • Multiple fixes to template construction scripts. Do not fall back on population average if initial template is specified but the file is not found. Check that all input images are successfully registered and that template updates complete successfully at each iteration.
  • Fix template prior cooking bug causing priors to be zeroed out due to precision errors.
  • Fix LesionFilling divide by zero error (#1258)
  • Fix bug in joint label fusion affecting multi-modality atlases (#1302)
  • Fix exception in PrintHeader spatial orientation reporting (#1382)
  • Denoise all inputs to antsAtroposN4.sh, and remove duplicate denoising steps to save time (#1373)

New features:

  • Option to pass an initial affine transform to antsBrainExtraction.sh.
  • Option to control SyN gradient steps in antsRegistrationSyN scripts.
  • Reproducible multi-threaded registration options (#1190)
  • Multiple updates to template building scripts. New options to control template averaging and sharpening, and make unbiased initialization more robust. Store intermediate templates after each iteration so that users can assess convergence and if needed, re-use previous templates for a new run of the script.
  • Kinematics tensors in ImageMath.
  • Support transforms of AFNI 5D statistical images.
  • Write HDF5 ITK transform objects with antsApplyTransforms (#1317).
  • Add gradient filter smoothing options to improve metric stability (#1351)
  • Allow smoothing in antsMotionCorr to have mm or vox units, change defaults to vox for consistency with antsRegistration (#1350).

Code and documentation:

  • Clarify behavior of N4 rescaling option.
  • Updates to template and cortical thickness script usage.
  • Improved VNL vector and matrix implementation.
  • Update code for C++14.
  • Add compilation support for IBM POWER architectures (#1243)
  • Update version information so that source tarballs contain a correct version number.
  • Move Docker CI build to Github Actions
  • Run short tests during Docker CI build (#1365).
  • Remove support for deprecated git protocol
  • Update ITK

What's Changed

New Contributors

Full Changelog: v2.3.5...v2.4.0