Zero-Shot Monocular Motion Segmentation in the Wild by Combining Deep Learning with Geometric Motion Model Fusion

arXiv:2405.01723v1 Announce Type: new
Abstract: Detecting and segmenting moving objects from a moving monocular camera is challenging in the presence of unknown camera motion, diverse object motions and complex scene structures. Most existing methods rely on a single motion cue to perform motion segmentation, which is usually insufficient when facing different complex environments. While a few recent deep learning based methods are able to combine multiple motion cues to achieve improved accuracy, they depend heavily on vast datasets and extensive …

abstract arxiv cs.ai cs.cv deep learning diverse fusion moving object objects segmentation type zero-shot