Unverified50% confidenceFactExact time
This project uses Meta SAM (Segment Anything Model) for object recognition.
1
Sources
50%
Confidence
Long-term
Relevance
10/30/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedSAM (Segment Anything Model) is a model open-sourced by Meta that achieves zero-shot segmentation of arbitrary objects through large-scale pre-training on 11 million images and 1 billion masks.79% similarVerifiedMeta's research division developed models such as SAM (Segment Anything Model) and ImageBind.76% similarUnverifiedSegment Anything（SAM）是Meta推出的通用图像分割基础模型，能够零样本分割任意物体71% similarUnverifiedSAM, developed by Meta, can perform zero-shot segmentation of any object in an image; its core innovation is framing segmentation as a promptable, interactive problem.70% similarUnverifiedMeta 的 Segment Anything Model（SAM）将高质量智能抠像能力完全开源化70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1116278API
curl https://kongchang.com/api/v1/knowledge/claims/1116278MCP
get_claim(id=1116278)