Abstract: In this paper, we introduce a simple but effective training-free pipeline for handling the task of text-to-video object segmentation. Our approach leverages open-source Multimodal Large ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results