A case for using rotation invariant features in state of the art feature matchers (CVPR 2022)

Описание: A case for using rotation invariant features in state of the art feature matchers, by Georg Bökman (Chalmers University of Technology), Fredrik Kahl (Chalmers)

The aim of this paper is to demonstrate that a state of the art feature matcher (LoFTR) can be made more robust to rotations by simply replacing the backbone CNN with a steerable CNN which is equivariant to translations and im- age rotations. It is experimentally shown that this boost is obtained without reducing performance on ordinary illumi- nation and viewpoint matching sequences.

CVPR2022 Image Matching Workshop
https://image-matching-workshop.githu...