Segment Anything | VitaVision
PaperICCV · 2023

Segment Anything

Alexander KirillovEric MintunNikhila RaviHanzi MaoChloe RollandLaura GustafsonTete XiaoSpencer WhiteheadAlexander C. BergWan‐Yen LoPiotr DollárRoss Girshick
arXiv 2304.02643DOIView in graph
1page built on it
1more page cites it
4narratives
4registry papers cite it

In the Atlas

ModelPrimary source of
SAM

Promptable segmentation foundation model family — SAM (v1, 2023) introduces image-prompt segmentation with a heavy ViT-H encoder and lightweight transformer decoder trained on the 1.1B-mask SA-1B dataset; SAM 2 (2024) extends to video via a streaming memory module on a Hiera hierarchical-ViT encoder; SAM 3 (2025) generalises from single-object prompts to *concept* prompts (free-form noun phrases or visual exemplars) via a presence token, segmenting all matching instances on images and videos.

Cited as a reference by 1 page, grouped by domain

Segmentation 1
MobileSAM

Appears in 4 narratives

NarrativeFoundation Models for VisionNarrativeSegmentation, from Energy to PromptNarrativeVision Under a Latency BudgetNarrativeWhere Did the Inductive Bias Go?

Lineage

Papers in the Atlas registry only · one dot per paper
Built upon by · 4
’04’05’06’07’08’09’10’11’12’13’14’15’16’17’18’19’20’21’22’23’24’25Faster Segment Anything: Towards Lightweight SAM for Mobile Applications (2023)MobileSAMv2: Faster Segment Anything to Everything (2023)SAM 2: Segment Anything in Images and Videos (2024)SAM 3: Segment Anything with Concepts (2025)Efficient Graph-Based Image Segmentation (2004)Fully Convolutional Networks for Semantic Segmentation (2015)U-Net: Convolutional Networks for Biomedical Image Segmentation (2015)Mask R-CNN (2017)DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs (2018)