SAM 2: Segment Anything in Images and Videos
Paper
• 2408.00714 • Published
• 120
Computer Vision
Do we still need a network for specific computer vision tasks anymore today?
Note 1. Process video frames one at a time, equipped with a memory attention module to attend to the previous memories of the target object.
-
facebook/sam2.1-hiera-large
Mask Generation • 0.2B • Updated • 222k • 129