Rui Zhai

Mitigating Task and Data Domain Discrepancy for Open- Vocabulary Video Instance Segmentation

Current Open-Vocabulary Video Instance Segmentation (OVVIS) approaches rely on vision–language models (VLMs) for generalization to unseen categories. However, they often fuse segmentation queries with VLM features in a simple and direct manner, overlooking the domain discrepancy between these two tasks. This mismatch l...