Project Page: http://www.cscvlab.com/research/SKS-Homography/
GitHub Page: https://github.com/cscvlab/SKS-Homography
Fast and Interpretable 2D Homography Decomposition: Similarity-Kernel-Similarity and Affine-Core-Affine Transformations. (Accepted by PAMI)
Previous four-point homography computation methods exhibit algebraic redundancy (arising from the construction of a sparse linear system) and geometric isolation (no link to other 2D primitives, transformation computations, and minimal vision problems). Our SKS and ACA methods offer distinct advantages in numerous aspects and demonstrate extreme efficiency.
We believe that SKS/ACA provides comprehensive solutions for plane-based geometric vision problems.
Furthermore, SKS/ACA holds promise for applications beyond its current scope, with potential extensions into fields such as linear algebra, projective geometry, and computer graphics.
