Method Details


Details for method 'ndnet'

 

Method overview

name ndnet
challenge pixel-level semantic labeling
details
publication Anonymous
project page / code
used Cityscapes data fine annotations
used external data
runtime 0.024 s
subsampling no
submission date February, 2019
previous submissions

 

Average results

Metric Value
IoU Classes 64.8348
iIoU Classes 34.7366
IoU Categories 84.4658
iIoU Categories 64.6744

 

Class results

Class IoU iIoU
road 97.7194 -
sidewalk 79.4887 -
building 89.0005 -
wall 39.587 -
fence 41.1877 -
pole 47.0118 -
traffic light 51.8359 -
traffic sign 58.692 -
vegetation 90.6959 -
terrain 66.848 -
sky 93.7165 -
person 74.5525 47.4073
rider 53.2526 24.0523
car 92.3632 83.9146
truck 48.2414 19.2039
bus 61.8825 26.4424
train 43.3018 18.6726
motorcycle 44.0309 18.9128
bicycle 58.4538 39.2865

 

Category results

Category IoU iIoU
flat 98.106 -
nature 90.4993 -
object 53.5964 -
sky 93.7165 -
construction 89.2107 -
human 74.706 48.4593
vehicle 91.4254 80.8894

 

Links

Download results as .csv file

Benchmark page