In vision-language models (VLMs), visual tokens usually consume a significant amount of computational overhead, despite their sparser information density compared to text tokens. To address this, ...

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 trying T-Rex demo link, which is ...

Abstract: Understanding and interpreting a script is essential for effective acting. Existing visualization methods, however, primarily focus on general narrative comprehension and often neglect ...

Visual grounding for remote sensing images (RSVG), as the frontier of the integration of computer vision and natural language processing technologies, aims to understand the content of input referring ...