Abstract: Existingstereo visual-inertial initialization methods, whether tightly or loosely coupled, rely critically on intermediate variables like feature correspondences and camera poses rather than ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results