Jul 29, 2024
Meta introduces a promptable model for identifying and tracking objects across images and video.
Open code and model checkpoints extend the Segment Anything approach from still images to moving scenes.
Explore on Humanity Unfolds
