1 paper
Nicolas Carion, Laura Gustafson, Yuan-Ting Hu +35
We present Segment Anything Model (SAM) 3, a unified model that detects, segments, and tracks objects in images and videos based on concept prompts, which we define as either short…