Currently we can use SIFT, SURF, ORB, BRISK and many other local features for image registration, but all of them are still having trouble in feature point discrimination in some cases (i.e. in different environments). I found that there are some studies have been conducted on using deep learning for local feature learning for registering images in specific domains such as medical images. What I need to know is whether there is any study on applying deep learning for learning a set of features which can be used for generic image registration.
Sebastian Zambanini
I would suggest to read the following recent publication: http://arxiv.org/abs/1512.09272
1 votes
0 thanks
- Badges
- Science topic
More Ravimal Bandara's questions See All
Similar questions and discussions