search

hkchengrex / Tracking-Anything-with-DEVA

[ICCV 2023] Tracking Anything with Decoupled Video Segmentation
https://hkchengrex.com/Tracking-Anything-with-DEVA/
Other
1.27k stars 129 forks source link

text prompt #69

Closed yangyang117 closed 8 months ago

yangyang117 commented 8 months ago

Hello, I would like to know if there are specific words for the names of certain items, such as slides, cups, etc? Is there any relevant document that can be used to search for text prompt?

hkchengrex commented 8 months ago

That depends on the image model. In our demo, that would be Grounded SAM https://github.com/IDEA-Research/Grounded-Segment-Anything