This is a minor release with some improvements.
Update `RegImage` to support more image formats as well as make it easy to add new formats
* add `RegImage` sub-classes for different file types
- `TiffFileRegImage` (.scn, .ndpi,.tiff,.tif) : uses `dask` + `zarr` to do memory-efficient computation of necessary data for registration / transformation
- `CziRegImage` (.czi) : Carl Zeiss image format, can perform read-time pre-processing like RGB -> greyscale or selection of individual channels to limit memory footprint
- `OmeTiffRegImage` (.ome.tiff,ome.tif): uses TiffFile to read images and parses OME metadata to get interleaved RGB information
- `MergeRegImage` (meta): used to transform multiple images' channels to a single OME-TIFF after registration if they output to the same size and data type (i.e. for cyclic IF)
- 'NpRegImage' (`np.ndarray`): Supports adding a registration image from a `numpy` array
- `SitkRegImage' (everything else): uses SimpleITK to read images as a last resort. Will read entire image into memory!
Add mask support and crop-to-mask support for registration
* support masks for registration
- Masks can be used in `elastix` to define pixels used in metric calculation
- add ability to automatically crop images based on associated masks's bounding box (can be useful if image dimensions differ greatly)
Add class to manage transformations
* use `RegTransform` class to manage transformations