
The author proposes a self-supervised method to jointly learn 3D motion and depth from monocular videos, addressing the limitations of existing methods in modeling dynamic scenes.


coremsg

self-supervised-learning-of-decomposed-object-aware-3d-motion-and-depth-from-monocular-videos


Self-supervised Learning of Decomposed Object-aware 3D Motion and Depth from Monocular Videos


title_rewrite


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.


motion-aware-loss-mal-enhancing-multi-frame-self-supervised-monocular-depth-estimation-in-dynamic-scenes-using-temporal-and-distillation-hints


Motion-Aware Loss (MAL): Enhancing Multi-Frame Self-Supervised Monocular Depth Estimation in Dynamic Scenes Using Temporal and Distillation Hints
