Abstract: State-of-the-art visual localization methods mostly rely on complex procedures to match local descriptors and 3D point clouds. However, these procedures can incur significant costs in terms ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results