I need to check how good the result of one image registration process' was. For it I thought of using Normalized Cross correlation between both images: base and result of the process. The problem is that I never worked with cross correlation and I dont know if it could be the better approach.
Then with the result I could see where the images are more dissimilar and try to improve the registration on this part of the image.
Would I need to make a two dimensional cross correlation? then I would obtain a map of dissimilarities,right?
Any help/suggestion would be useful