Has anyone tried SuperGlue for multi-modal feature matching? E.g., a depth image and RGB image?

[jediofgever]

No description provided.

[jediofgever]

I came accross this paper which apprently uses super glue on LIDAR intensity image and grayscale camera image.
https://arxiv.org/abs/2302.05094

[ashu1069]

I am working on training superglue for thermal images and eventually I have to use separately trained superglue for feature matching in RGB and thermal. Even the pre-trained one works quite well on RGB & thermal, however, there were inconsistencies.