Model docs ↗
The original SAM image-segmentation model extracts masks from point or box prompts, or automatically segments an image.
TOOLORA / Local reading edition
Segment Anything (SAM)
The original SAM image-segmentation model extracts masks from point or box prompts, or automatically segments an image.
Overview
Segment Anything is Meta's image-segmentation model, using prompts such as points and boxes to identify regions. It produces object masks, and the official code also provides automatic mask generation. Inference code, model weights and example notebooks support integration into custom image-processing workflows.
- Company country / region
- United States · Company headquarters
Access options
Official sources and references
- Official websitehttps://github.com/facebookresearch/segment-anything
- Documentationhttps://github.com/facebookresearch/segment-anything/blob/main/notebooks/predictor_example.ipynb
- Documentationhttps://github.com/facebookresearch/segment-anything/blob/main/notebooks/automatic_mask_generator_example.ipynb
Descriptions summarize public information; consult official websites for features, pricing and regional availability.