Our lab is currently using DIPY’s symmetric diffeomorphic registration to extract deformation maps from two registered images. I currently have these deformation maps, but no longer have other latent data. DIPY’s documentation makes it very easy and clear how to extract these maps (ie. get_forward_field, get_backward_field). What I’m interested in is given the deformation map, how can I “warp” one of my image to look like the other registered image? Would someone with experience with image co-registration and/or the original authors be able to comment?
Thank you!