Abstract: Visual place recognition (VPR) using deep networks has achieved state-of-the-art performance. However, most of them require a training set with ground truth sensor poses to obtain positive ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback