This dataset is a set of additional annotations for PASCAL VOC 2010. It goes beyond the original PASCAL semantic segmentation task by providing annotations for the whole scene. The statistics section has a full list of 400+ labels.

Below are some example segmentations from the dataset.

Below are some example class masks.

computer
sofa
tree

Training and Validation Set

The following download contains only the annotations for the training/validation set and a text file (labels.txt) containing the integer-to-name mapping.

trainval.tar.gz (30.7 MB)

  • MD5: df034edb2c12aa7d33b42b20bb1796e3
  • SHA1: e1375777199dcc02097e68aea755415ca173aa2d

Original images for the dataset can be downloaded from the PASCAL VOC 2010 website.

Testing Set

Please download the images from the PASCAL VOC website.

Project Specific Downloads

The segmentation results for the 33 categories and the label-to-name mapping can be downloaded here (trained on the train subset and tested on the val subset). This is the result of a 33-way classifier. These segmentations were used to train contextDPM in the paper.

33_context_labels.tar.gz (27.1 MB)

33_labels.txt (345 B)

The segmentation results for the 59 categories and the label-to-name mapping can be downloaded here (trained on the train subset and tested on the val subset). This is the result of a 59-way classifier. The accuracies in Tables 1 and 2 in the CVPR paper are computed according to these predictions.

59_context_labels.tar.gz (14.1 MB)

59_labels.txt (613 B)

back to top

Since the dataset is an annotation of PASCAL VOC 2010, it has the same statistics as those of the original dataset. Training and validation contains 10,103 images while testing contains 9,637 images.

Click on the panel below to expand the full class list. The following image count and average area are calculated only over the training and validation set. When the testing set is released these numbers will be updated. Image counts below may be zero because a class was present in the testing set but not the training and validation set.

Class Name# of ImagesAverage Area (% of image)
empty0
accordion25.26%
aeroplane59717.41%
air conditioner52.68%
antenna20.80%
artillery13.36%
ashtray181.11%
atrium113.64%
baby carriage1312.28%
bag3193.18%
ball252.12%
balloon182.88%
bamboo weaving112.78%
barrel33.58%
baseball bat11.05%
basket895.70%
basketball backboard65.37%
bathtub536.08%
bed12028.08%
bedclothes29627.01%
beer12.41%
bell21.24%
bench1195.49%
bicycle50317.20%
binoculars0
bird66412.36%
bird cage2144.05%
bird feeder1418.47%
bird nest621.69%
blackboard212.70%
board1379.67%
boat46616.67%
bone32.55%
book4163.80%
bottle8217.35%
bottle opener10.61%
bowl1372.57%
box4345.16%
bracelet10.76%
brick94.38%
bridge3110.39%
broom56.75%
brush12010.28%
bucket1752.73%
building306420.22%
bus39932.86%
cabinet53613.62%
cabinet door15.84%
cage2932.76%
cake28.76%
calculator11.30%
calendar153.30%
camel0
camera521.80%
camera lens0
can782.24%
candle431.04%
candle holder290.93%
cap163.45%
car127319.45%
card113.41%
cart116.56%
case914.12%
casette recorder11.82%
cash register10.88%
cat100633.30%
cd453.33%
cd player11.35%
ceiling58610.00%
cell phone430.49%
cello415.31%
chain101.52%
chair135711.87%
chessboard21.37%
chicken0
chopstick81.12%
clip10.47%
clippers29.44%
clock610.99%
closet127.69%
cloth8778.96%
clothes tree0
coffee40.59%
coffee machine21.41%
comb30.80%
computer1188.21%
concrete96.79%
cone301.32%
container614.34%
control booth0
controller34.67%
cooker0
copying machine0
coral0
cork0
corkscrew0
counter3111.19%
court214.94%
cow25622.98%
crabstick12.44%
crane120.56%
crate20.74%
cross30.23%
crutch20.52%
cup4322.46%
curtain38111.51%
cushion7512.60%
cutting board32.50%
dais110.10%
disc22.49%
disc case164.96%
dishwasher0
dock310.26%
dog120425.23%
dolphin0
door45510.01%
drainer10.09%
dray110.38%
drink dispenser11.92%
drinking machine17.36%
drop0
drug11.52%
drum98.77%
drum kit87.93%
duck158.21%
dumbbell0
earphone81.21%
earrings11.07%
egg36.92%
electric fan112.57%
electric iron22.07%
electric pot12.02%
electric saw13.29%
electronic keyboard0
engine23.94%
envelope0
equipment277.52%
escalator0
exhibition booth217.59%
extinguisher60.95%
eyeglass152.72%
fan402.05%
faucet42.27%
fax machine12.58%
fence102110.79%
ferris wheel110.09%
fire extinguisher20.75%
fire hydrant24.74%
fire place419.53%
fish74.60%
fish tank325.04%
fishbowl118.00%
fishing net17.74%
fishing pole10.21%
flag1002.05%
flagstaff31.58%
flame0
flashlight20.38%
floor189516.07%
flower1907.67%
fly10.43%
foam0
food2855.16%
footbridge110.82%
forceps0
fork1040.63%
forklift0
fountain11.40%
fox0
frame175.86%
fridge6910.07%
frog10.70%
fruit20.77%
funnel10.37%
furnace231.60%
game controller40.66%
game machine313.31%
gas cylinder0
gas hood0
gas stove0
gift box210.96%
glass215.24%
glass marble18.14%
globe0
glove64.34%
goal410.04%
grandstand3018.55%
grass274223.28%
gravestone0
ground298720.98%
guardrail356.09%
guitar534.83%
gun21.07%
hammer10.08%
hand cart114.87%
handle24.16%
handrail73.75%
hanger30.66%
hard disk drive0
hat274.07%
hay321.08%
headphone70.49%
heater48.10%
helicopter312.77%
helmet622.32%
holder22.23%
hook20.61%
horse42922.31%
horse-drawn carriage2512.70%
hot-air balloon184.05%
hydrovalve0
ice631.35%
inflator pump0
ipod150.75%
iron22.15%
ironing board19.63%
jar182.80%
kart116.90%
kettle11.96%
key41.86%
keyboard1404.69%
kitchen range127.54%
kite10.26%
knife920.61%
knife block20.30%
ladder123.91%
ladder truck10.84%
ladle12.83%
laptop7510.55%
leaves733.39%
lid11.37%
life buoy214.83%
light9081.18%
light bulb10.68%
lighter50.21%
line10.14%
lion0
lobster0
lock10.15%
machine0
mailbox30.39%
mannequin111.90%
map332.42%
mask0
mat3511.11%
match book12.29%
mattress31.72%
menu244.34%
metal2164.28%
meter box13.43%
microphone440.71%
microwave263.61%
mirror815.60%
missile15.71%
model127.81%
money16.08%
monkey12.75%
mop30.66%
motorbike47124.63%
mountain72810.75%
mouse1220.43%
mouse pad503.58%
musical instrument711.16%
napkin1202.60%
net36.81%
newspaper303.99%
oar105.48%
ornament20.79%
outlet260.48%
oven810.24%
oxygen bottle14.21%
pack2873.18%
pan103.51%
paper2984.02%
paper box40.99%
paper cutter66.75%
parachute43.45%
parasol174.14%
parterre114.77%
patio123.16%
pelage0
pen440.33%
pen container81.14%
pencil20.65%
person391920.64%
photo32.69%
piano48.83%
picture6745.70%
pig11.41%
pillar83.16%
pillow5511.18%
pipe162.01%
pitcher33.36%
plant10614.52%
plastic116.37%
plate3464.44%
platform15415.39%
player242.37%
playground15.70%
pliers10.39%
plume11.76%
poker10.73%
poker chip17.81%
pole15551.53%
pool table129.81%
postcard11.72%
poster1856.82%
pot3103.08%
pottedplant62511.89%
printer342.92%
projector20.52%
pumpkin27.72%
rabbit10.14%
racket14.75%
radiator273.93%
radio62.22%
rail316.12%
rake10.26%
ramp23.15%
range hood83.98%
receiver131.12%
recorder223.21%
recreational machines161.90%
remote control670.93%
road110221.97%
robot0
rock39610.72%
rocket0
rocking horse10.74%
rope5071.35%
rug18120.69%
ruler11.61%
runway211.72%
saddle851.95%
sand4636.43%
saw11.53%
scale23.64%
scanner11.43%
scissors50.72%
scoop10.04%
screen413.73%
screwdriver20.32%
sculpture737.67%
scythe0
sewer226.70%
sewing machine31.06%
shed22.25%
sheep29721.95%
shell32.43%
shelves4487.64%
shoe742.68%
shopping cart24.54%
shovel21.07%
sidecar215.60%
sidewalk3828.65%
sign5922.32%
signal light430.98%
sink2415.51%
skateboard10.81%
ski35.71%
sky323126.77%
sled24.28%
slippers14.63%
smoke5610.42%
snail0
snake315.25%
snow17328.71%
snowmobiles116.52%
sofa83125.31%
spanner0
spatula11.36%
speaker1183.48%
speed bump0
spice container230.99%
spoon750.51%
sprayer0
squirrel16.33%
stage1649.30%
stair3621.55%
stapler30.79%
stick40.28%
sticky note21.64%
stone0
stool93.18%
stove102.22%
straw60.47%
stretcher0
sun13.99%
sunglass40.86%
sunshade35.83%
surveillance camera10.16%
swan10.22%
sweeper0
swim ring29.90%
swimming pool13.01%
swing12.09%
switch50.51%
table15669.76%
tableware34.19%
tank22.63%
tap122.30%
tape90.51%
tarp243.12%
telephone711.72%
telephone booth0
tent3612.26%
tire353.29%
toaster51.70%
toilet212.68%
tong11.80%
tool33.06%
toothbrush10.80%
towel318.69%
toy3565.21%
toy car230.91%
track3569.15%
train46528.46%
trampoline115.35%
trash bin913.03%
tray95.46%
tree323218.02%
tricycle0
tripod28.27%
trophy21.04%
truck1266.39%
tube33.82%
turtle10.98%
tvmonitor53615.32%
tweezers0
typewriter13.40%
umbrella134.32%
unknown37348.99%
vacuum cleaner64.15%
vending machine223.66%
video camera111.94%
video game console111.77%
video player372.97%
video tape34.09%
violin44.37%
wakeboard0
wall279422.23%
wallet71.26%
wardrobe218.93%
washing machine1715.58%
watch60.81%
water89932.68%
water dispenser53.52%
water pipe32.94%
water skate board10.60%
watermelon27.00%
whale0
wharf0
wheel817.21%
wheelchair53.78%
window75611.07%
window blinds11.57%
wineglass1972.95%
wire1131.17%
wood4628.29%
wool24.78%
close class list

back to top

The classes are not drawn from a fixed pool. Instead labelers were free to either select or type in what they believe to be the appropriate class and to determine what the appropriate object granularity is. We decided to merge/split some of the categories so the current number of categories is different from what we mentioned in the CVPR 2014 paper.

When using this dataset it is important that you examine classes to ensure they match your intended use. For example, sand is often labeled independently despite also being considered ground. Those interested in ground may want to cluster sand and ground together along with other classes.

back to top

back to top

We thank Viet Nguyen for coordinating and leading the efforts for cleaning up the annotations. We would like to acknowledge the support by Implementation of Technologies for Identification, Behavior, and Location of Human based on Sensor Network Fusion Program through the Korean Ministry of Trade, Industry and Energy (Grant Number: 10041629). We would also like to thank National Science Foundation for grant 1317376 (Visual Cortex on Silicon. NSF Expedition in Computing).

back to top