REMM:Rotation-Equivariant Framework for End-to-End Multimodal Image Matching
The objective of this research is Rotation-Equivariant Framework for End-to-End Multimodal Image Matching.
Pretrained weights will be able soon...
(a)OPT-SAR：SEN2 & SEN1
(b)OPT-SAR：Google Earth & GF-3
(c)OPT-NIR: Aerial image & Aerial image
(d)OPT-NIR: Camera & Camera
https://arxiv.org/abs/2407.11637