Language-Guided Instance-Aware Domain-Adaptive Panoptic Segmentation

arXiv:2404.03799v1 Announce Type: new
Abstract: The increasing relevance of panoptic segmentation is tied to the advancements in autonomous driving and AR/VR applications. However, the deployment of such models has been limited due to the expensive nature of dense data annotation, giving rise to unsupervised domain adaptation (UDA). A key challenge in panoptic UDA is reducing the domain gap between a labeled source and an unlabeled target domain while harmonizing the subtasks of semantic and instance segmentation to limit catastrophic interference. …

abstract annotation applications arxiv autonomous autonomous driving challenge cs.ai cs.cv data data annotation deployment domain domain adaptation driving giving however instance key language nature panoptic segmentation segmentation type unsupervised