SAM 1: selecting an object with a point or a box

The human problem

Drawing precise boundaries around objects is slow, repetitive work in image annotation and analysis.

The prior constraint

Segmentation systems usually needed category-specific labels and training before they could help on a new collection.

AI’s actual role

A promptable vision model produces candidate masks for objects without training a new model for each object category.

The documented result

Meta released SAM 1 inference code, Apache-2.0 model checkpoints, notebooks, and separate access to the SA-1B dataset in April 2023.

Why it may matter

Promptable masks can move human effort from drawing every boundary to checking and correcting proposals.

Limitations

A plausible mask may still be wrong, especially in unfamiliar imagery. Apache-2.0 model artifacts do not settle the rights attached to input images, dataset access, or downstream uses.

This is the historical SAM 1 release, not a claim about later Segment Anything generations. The official repository establishes Apache-2.0 model artifacts and intended workflows; SA-1B data access and downstream image rights remain separate.

Unresolved questions