|
gptkbp:instanceOf
|
gptkb:model
|
|
gptkbp:abbreviation
|
gptkb:SAM
|
|
gptkbp:application
|
autonomous vehicles
computer vision
robotics
medical imaging
|
|
gptkbp:architecture
|
transformer-based
|
|
gptkbp:arxivID
|
2304.02643
|
|
gptkbp:author
|
gptkb:Alexander_Kirillov
gptkb:Hanzi_Mao
Cheng-Yang Fu
Eric Mintun
Nikhila Ravi
and others
|
|
gptkbp:citation
|
high (hundreds+ as of 2024)
|
|
gptkbp:developedBy
|
gptkb:Meta_AI
|
|
gptkbp:fullName
|
gptkb:Segment_Anything_Model
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
Segment Anything
|
|
gptkbp:influenced
|
MedSAM
SAM-Adapter
segment-anything-3d
|
|
gptkbp:input
|
gptkb:illustrator
|
|
gptkbp:inputPromptType
|
boxes
masks
points
text (limited)
|
|
gptkbp:license
|
gptkb:Apache_License_2.0
|
|
gptkbp:memiliki_tugas
|
image segmentation
|
|
gptkbp:notable_for
|
gptkb:Hugging_Face_Spaces
segment-anything.com
|
|
gptkbp:notableFeature
|
scalable to large datasets
interactive segmentation
promptable segmentation
works on any object, any image
zero-shot generalization
|
|
gptkbp:notablePublication
|
gptkb:Segment_Anything
https://arxiv.org/abs/2304.02643
|
|
gptkbp:openSource
|
true
|
|
gptkbp:output
|
segmentation mask
|
|
gptkbp:relatedTo
|
gptkb:CLIP
gptkb:Mask_R-CNN
ViT (Vision Transformer)
|
|
gptkbp:releaseYear
|
2023
|
|
gptkbp:repository
|
https://github.com/facebookresearch/segment-anything
|
|
gptkbp:SA-1B_datasetSize
|
1 billion masks
|
|
gptkbp:trainer
|
gptkb:SA-1B_dataset
|
|
gptkbp:bfsParent
|
gptkb:Segment_Anything_Model
|
|
gptkbp:bfsLayer
|
6
|