prajwel/aafitrans: v0.2.1

Aafitrans is a Python package that builds upon the capabilities of the Astroalign package's find_transform function. It incorporates several modifications to improve its functionality and performance:

1. Enhanced RANSAC Algorithm: The RANSAC algorithm used in Aafitrans has been optimized to provide a solution that minimizes the sum of squared residuals. This improvement ensures a more accurate transformation estimation.

2. Arun and Horn's Method: Aafitrans replaces Umeyama's method from scikit-image with Arun and Horn's method for estimating 'euclidean' or 'similarity' transformations.

3. Reflection Support: Unlike Astroalign, Aafitrans enables the matching of coordinate lists that include reflection along one axis. This enhancement expands the range of transformations that can be accurately estimated.

4. Extended Transformation Options: Aafitrans supports all transformations available in the scikit-image library, providing a comprehensive set of options for aligning and transforming images.

Note: only 'euclidean', 'similarity', and 'affine' transformations have been tested.

1. Improved Matching Efficiency: The kdtree_search_radius parameter in Aafitrans allows users to set the search radius for matches, enabling faster and more efficient matching of corresponding points between images.

2. Reproducible Results: Aafitrans introduces the seed parameter, which can be set during each run to ensure the reproducibility of results. This feature is particularly useful for research and debugging purposes.

3. Dependency Optimization: Aafitrans eliminates the need for the sep and bottleneck packages as dependencies, streamlining the installation process and reducing potential compatibility issues.