Kernekoncepter
This research proposes Motion-Aware Loss (MAL), a novel plug-and-play module that leverages temporal coherence and an enhanced distillation scheme to improve the accuracy of multi-frame self-supervised monocular depth estimation, particularly in dynamic scenes.
Statistik
Adding MAL into previous state-of-the-art methods leads to a reduction in depth estimation errors by up to 4.2% and 10.8% on KITTI and CityScapes benchmarks, respectively.
In the KITTI dataset, dynamic category objects account for only 0.34% of the pixels.