Method Details


Details for method 'Fast OCNet'

 

Method overview

name Fast OCNet
challenge pixel-level semantic labeling
details
publication Anonymous
project page / code
used Cityscapes data fine annotations
used external data ImageNet
runtime n/a
subsampling no
submission date January, 2019
previous submissions

 

Average results

Metric Value
IoU Classes 82.0642
iIoU Classes 61.0092
IoU Categories 91.6521
iIoU Categories 80.6902

 

Class results

Class IoU iIoU
road 98.7381 -
sidewalk 87.3075 -
building 93.6535 -
wall 59.856 -
fence 62.5746 -
pole 70.4883 -
traffic light 78.3743 -
traffic sign 80.9792 -
vegetation 93.8269 -
terrain 73.4229 -
sky 95.5608 -
person 87.6319 71.613
rider 75.0614 54.3016
car 96.371 91.0479
truck 75.8239 43.3191
bus 90.0985 57.5461
train 88.8693 54.45
motorcycle 71.9339 51.8007
bicycle 78.6485 63.9951

 

Category results

Category IoU iIoU
flat 98.7571 -
nature 93.5265 -
object 76.1714 -
sky 95.5608 -
construction 94.0175 -
human 87.6177 72.3816
vehicle 95.9141 88.9989

 

Links

Download results as .csv file

Benchmark page