Features
Below is a list of features which are either already implemented or are on the roadmap.
Dimensionality
- 2D
- 3D
Registration
Pairwise registration methods
- Phase correlation
- ANTsPy
- Elastix (
itk-elastix) - Bead alignment
- Phase correlation for rotation + translation
Global paramater resolution
- Graph construction
- Automatic determination of reference view
- Parameter concatenation along graph connectivity paths
- Global optimization of registration parameters from (potentially overdetermined) pairwise transforms
Transformations
- Chunked
dask_image.ndinterp.affine_transform - Cupy-based transform
- Chaining transformations instead of working with static coordinate systems
Fusion
General
- Modular API to plug in different fusion and weight functions
- Support for fusion label maps
- Cupy-based fusion
Supported fusion methods:
- Weighted average
- Maximum intensity projection
- Multi-view deconvolution
Supported weights:
- Linear blending
- Content-based
Supported data formats
- OME-Zarr
- Zarr based intermediate file format for reading and writing, compatible with parallel dask workflows: multiscale-spatial-data
- CZI input
- Multi-positioning
- Light-sheet
- TIF input
- TIF writing
Visualization
Napari
See napari-stitcher.
- 2D slice view: multiscale rendering
- 3D rendered view:
- Lowest scale
- Chunked rendering
- Colormaps optimized for highlighting differences between overlapping views
Supported usage modes
- As a library to build custom reconstruction workflows
- Napari plugin
- Convenience function for processing on HPC