Method Details


Details for method 'Seg_UJS'

 

Method overview

name Seg_UJS
challenge pixel-level semantic labeling
details Previously listed as "WRnet".
publication Anonymous
project page / code
used Cityscapes data fine annotations
used external data ImageNet
runtime n/a
subsampling no
submission date July, 2020
previous submissions 1, 2

 

Average results

Metric Value
IoU Classes 80.4497
iIoU Classes 56.5501
IoU Categories 91.2834
iIoU Categories 80.0296

 

Class results

Class IoU iIoU
road 98.5935 -
sidewalk 86.4153 -
building 93.5509 -
wall 65.011 -
fence 62.3185 -
pole 69.1139 -
traffic light 75.9747 -
traffic sign 79.5924 -
vegetation 93.7551 -
terrain 72.7253 -
sky 95.6226 -
person 86.7818 71.0455
rider 71.2256 47.8503
car 95.9022 89.7471
truck 71.8294 37.1893
bus 86.1011 51.9883
train 79.4808 45.3433
motorcycle 68.5196 46.3543
bicycle 76.0315 62.8826

 

Category results

Category IoU iIoU
flat 98.7092 -
nature 93.5131 -
object 75.0182 -
sky 95.6226 -
construction 93.7636 -
human 86.895 72.1955
vehicle 95.4621 87.8636

 

Links

Download results as .csv file

Benchmark page