ML//Multimodal//SAM

- Segment Anything Model (Meta, 2023): foundation model for image segmentation.

Segment Anything Model (Meta, 2023): foundation model for image segmentation.

Prompt with points, boxes, or text → get pixel-precise masks for any object.

SAM 2 extends to video — track and segment objects across frames.

GroundingDINO + SAM: text detection + pixel segmentation for open-vocabulary understanding.