Boosting Feature Matcher Performance

Home

Discover

Bookmarks

Keep Up With Top Trending Papers In CS / AI / ML ...

Source arXiv

Computer Vision

Boosting Feature Matcher Performance

A case for using rotation invariant features in state of the art feature matchers

A state of the art featurematcher (loftr) can be made more robust to rotations by simply replacing the backbone cnn with a steerable steerable cnn which is equivariant to translations and image rotations.

It is experimentally shown that this boost is obtained without reducing performance on ordinary illumination and viewpoint matching sequences.

Authors

Georg Bökman, Fredrik Kahl

◐ Latest Activity

◐ Recommended Members