NUS-WIDE-OBJECT
The Descriptions of the Files Included in the NUS-WIDE-OBJECT
The
NUS-WIDE-OBJECT is a real world object image dataset. This dataset is intended
for several object-based tasks, such as object categorization, object based
image retrieval, image annotations, etc. As a subset of NUS-WIDE, it consists
of 30 object categories and 30,000 images in total. It has 17,927 images for
training and 12,073 images for testing image. The figure below shows the list of
object categories and their training and testing image numbers
respectively.
Total
unique number of images in database_object: 30,000
number
of images with multiple labels in database_object: 6,047
total
training image: 17927 total testing image: 12073
category
bear - train 330 - test 251
category
birds - train 1012 - test 682
category
boats - train 1352 - test 843
category
book - train 94 - test 62
category
cars - train 677 - test 438
category
cat - train 623 - test 423
category
computer - train 140 - test 109
category
coral - train 731 - test 445
category
cow - train 232 - test 169
category
dog - train 682 - test 483
category
elk - train 234 - test 140
category
fish - train 728 - test 469
category
flags - train 99 - test 71
category
flowers - train 2,400 - test 1,550
category
fox - train 194 - test 144
category
horses - train 482 - test 305
category
leaf - train 675 - test 460
category
plane - train 839 - test 578
category
rocks - train 1,812 - test 1,203
category
sand - train 656 - test 456
category
sign - train 363 - test 261
category
statue - train 262 - test 165
category
sun - train 1,138 - test 810
category
tiger - train 194 - test 118
category
tower - train 707 - test 508
category
toy - train 779 - test 523
category
train - train 338 - test 244
category
tree - train 1,504 - test 1,024
category
vehicle - train 2,205 - test 1,468
category
whales - train 135 - test 91
category zebra - train 92 - test 53
-------------------------
raw data
-----------------------------
NUS-WIDE-OBJECT
concept list
bear
birds
boats
book
cars
cat
computer
coral
cow
dog
elk
fish
flags
flowers
fox
horses
leaf
plane
rocks
sand
sign
statue
sun
tiger
tower
toy
train
tree
vehicle
whales
zebra
If
you have any questions about the NUS-WIDE dataset, please contact Dr. Jinhui Tang: