In this work, we introduce DINOv, a Visual In-Context Prompting framework for referring and generic segmentation tasks. For visualization and demos, we also recommend ...

Abstract: One-shot medical image segmentation (MIS) aims to cope with the expensive, time-consuming, and inherent human bias annotations. One prevalent method to address one-shot MIS is joint ...