BDD100K dataset

GPTKB entity

Statements (52)
Predicate Object
gptkbp:instanceOf gptkb:dataset
gptkbp:annotationType attributes
labels
polygons
bounding boxes
instance masks
tracking IDs
gptkbp:citation Yu, Fisher, et al. 'BDD100K: A Diverse Driving Video Database with Scalable Annotation Tooling.' arXiv preprint arXiv:1805.04687 (2018).
gptkbp:contains video sequences
100,000 driving videos
annotated images
domain adaptation labels
drivable area labels
instance segmentation labels
lane marking labels
multi-object tracking labels
object detection labels
semantic segmentation labels
gptkbp:createdBy gptkb:Berkeley_Artificial_Intelligence_Research_(BAIR)
gptkbp:domain autonomous driving
gptkbp:durationPerVideo 40 seconds
gptkbp:format gptkb:DVD
gptkb:JSON
images
annotations
gptkbp:fullName Berkeley DeepDrive 100K dataset
https://www.w3.org/2000/01/rdf-schema#label BDD100K dataset
gptkbp:license Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International
gptkbp:numberOfRooms 100,000
gptkbp:numberOfStudents 10 (object detection)
40 (semantic segmentation)
gptkbp:numberOfVideos 100,000
gptkbp:refreshRate 30 fps
gptkbp:relatedTo gptkb:Cityscapes_dataset
KITTI dataset
nuScenes dataset
gptkbp:releaseYear 2018
gptkbp:resolution 1280x720 pixels
gptkbp:usedFor object detection
semantic segmentation
computer vision research
scene understanding
instance segmentation
domain adaptation
lane detection
autonomous vehicle perception
drivable area estimation
multi-object tracking
gptkbp:website https://bdd-data.berkeley.edu/
gptkbp:分布地区 gptkb:United_States
gptkbp:bfsParent gptkb:Berkeley_DeepDrive
gptkbp:bfsLayer 7