Method Details


Details for method 'SAITv2'

 

Method overview

name SAITv2
challenge pixel-level semantic labeling
details
publication Anonymous
project page / code
used Cityscapes data fine annotations, coarse annotations
used external data ImageNet
runtime 0.005 s
GPU TITAN X
subsampling no
submission date April, 2018
previous submissions

 

Average results

Metric Value
IoU Classes 75.7279
iIoU Classes 49.1373
IoU Categories 88.7936
iIoU Categories 72.896

 

Class results

Class IoU iIoU
road 98.4236 -
sidewalk 84.8901 -
building 92.1486 -
wall 58.4748 -
fence 55.1549 -
pole 60.1606 -
traffic light 66.7951 -
traffic sign 72.6596 -
vegetation 92.8146 -
terrain 70.7514 -
sky 95.1077 -
person 81.6379 59.3891
rider 63.483 38.6125
car 94.9539 87.1149
truck 69.4886 32.6753
bus 81.5235 46.2238
train 71.323 42.1448
motorcycle 58.8096 34.4192
bicycle 70.2292 52.5186

 

Category results

Category IoU iIoU
flat 98.5201 -
nature 92.5059 -
object 67.0913 -
sky 95.1077 -
construction 92.3252 -
human 81.7865 60.8585
vehicle 94.2187 84.9335

 

Links

Download results as .csv file

Benchmark page