Labeled databases for object detection (original) (raw)
Name
Objects
Background
Annotation
#ex per class
~500 categories (indoor and outdoor objects)
~3000 labeled natural images
Bounding polygons
50-500
107 categories (indoor and outdoor objects)
Natural images
Bounding polygons
~200
Side views of cars
Natural images
Bounding boxes
~500
Motorbikes, cars, cows
Natural images
Bboxes + seg masks
~100
Bicycles, cars, cows, pedestirans
Natural images
Segmentation masks
~300
Planes, motorbikes, faces, cars.
Close up photos
Bounding boxes
500-1000
101 categories
Close up photos
None
~50-200
Cows, bicycles, people, etc. (meta site)
Natural images
Various
Various
Chairs, bicycles, etc.
Natural images
Labeled edges
~100
100s
Natural images
Captions
?
10s
Corel segmented with Ncuts
Captions
?
Frontal faces, cars, pedestrians
None
Isolated patches
~800
Toy cars, animals, etc (many angles)
Synthetic cluttered backgrounds
Isolated patches
~10
Frontal faces
None
Isolated patches
~1000
Frontal faces (meta site)
None
Isolated patches
?