Skip to content

Evaluation

Yue Pan edited this page Sep 24, 2021 · 4 revisions

Evaluation metrics:

ATE, ARE defined in KITTI Odometry benchmark

Evaluation results:

The open-sourced version of MULLS in this repository performs a bit worse than the reported one with an average ATE of 0.55% on KITTI seq00-10.

KITTI online leaderboard (seq.11-21)