Hi,
I'd like to know how accuracy the mono-camera could be. In the paper, the ATE evaluation metrics compute the total trajectory error.
However, it's not the official evaluation metrics of KITTI odometry DataSet.
In addition, I'd like to know the reason why the rotation alignment of reconstructed path is needed when computing the ATE?
Hi, I'd like to know how accuracy the mono-camera could be. In the paper, the ATE evaluation metrics compute the total trajectory error. However, it's not the official evaluation metrics of KITTI odometry DataSet. In addition, I'd like to know the reason why the rotation alignment of reconstructed path is needed when computing the ATE?