Abstract: Learning-based visual relocalizers exhibit leading pose accuracy, but require hours or days of training. Since training needs to happen on each new scene again, long training times make ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results