NEWS

rpyANTs 0.0.7 (2026-10-12)

YAELPreprocess.get_template_mapping no longer fails on transform files without from/to entities, matches the subject code case-insensitively when it is spelled one way, and uses transforms whose sub entity differs from the subject code (renamed or copied folders) only when they all come from the subject under which the native image was imported
Restored the tissue probability maps and additional image modalities as extra metrics in the final normalization stage; they were not used since version 0.0.6
Fixed image types starting with postop (for example postopT1w) being saved with the session label ses-preop instead of ses-postop by YAEL; images saved by earlier versions are still found
Fixed the final registration not being cached when normalizing without antspynet; its transform files are now stored in the working directory, and a working directory that is used again returns the stored registration

Improved normalization for irregular brain
Fixed issues where atlases in nii are ignored
Added extract_coregistartion to extract index to world and world to world matrix
Added supports to the symmetric MNI152 templates
Normalization allows with_skull or without skulls; cost_function can be customized too
Users have choices to disable antspynet for consistency
Suppressed a test on MacOS due to ImageIO issue

Added backward compatibility support for Py39
Fixed the missing pow function from numpy
Added atropos segments for templates
New normalization: added atropos segments for normalization to templates
ANTsPyNet model locations are stored in package data directory to comply with CRAN policy
Always run abp during normalization
Added brain extraction function using base ANTs
Fixed issue when subject code contains underscore, causing non-comply BIDS format, causing parsing errors
Added fsaverage template for normalization

Added get_xform and set_xform to get and set qform in ANTsImage
Removed dependence nibabel, which might conflict with ANTsPy on certain platforms
Using the full path to ANTsImage class instead of the shortcut one from Python to avoid issues where ANTsImage is not exported (ANTsPyx==0.5.3)
Using multiple imaging files and faster normalization parameters when normalizing the native images to template
Fixed get_lib_fn not available under ANTsPyx>=0.5.3
Changed default Python version
Fixed bugs when mapping native ROI to template
Supported YAEL processing pipeline and use ANTsImage as inputs
Added mapping images from native to template
Fixed a typo in folder naming
Method get_native_mapping gets correct absolute paths
Allowed YAEL class to add default images through constructor
Added intensity correction
Exported ensure_template and added extra information
Set time-out to ensure downloading the correct templates
Added YAEL class to support pipelines mentioned in the paper for image registration and normalization

Added registration code that reproduces Halpern lab results
Added t1_preprocess to align MRI to template using non-linear registration
No more injection to Python on Windows as the memory pointer crash has been fixed by upstream developers
No active binding to antspynet as this Python package might fail to load on some systems

Added antspynet to the installation
Added ants_apply_transforms and ants_apply_transforms_to_points to apply registration transforms
Added ants_resample_image to resample images
Added antspynet_brain_extraction to extract brain and strip skulls
Added antspynet_deep_atropos and antspynet_desikan_killiany_tourville_labeling to segment brain images