Method Details


Details for method 'MiniNet-v2-coarse'

 

Method overview

name MiniNet-v2-coarse
challenge pixel-level semantic labeling
details
publication Anonymous
project page / code
used Cityscapes data fine annotations, coarse annotations
used external data
runtime 0.012 s
gpu titan xp
subsampling 2
submission date February, 2019
previous submissions

 

Average results

Metric Value
IoU Classes 67.7527
iIoU Classes 39.2439
IoU Categories 86.1161
iIoU Categories 68.3054

 

Class results

Class IoU iIoU
road 97.9668 -
sidewalk 81.538 -
building 89.9785 -
wall 44.5724 -
fence 44.1352 -
pole 54.5927 -
traffic light 57.7441 -
traffic sign 63.9148 -
vegetation 91.6414 -
terrain 68.1583 -
sky 94.5225 -
person 75.4345 52.8906
rider 54.5245 27.7511
car 93.1794 84.7273
truck 50.1238 19.6245
bus 66.8207 33.95
train 53.1031 30.6134
motorcycle 44.0468 23.0617
bicycle 61.3028 41.3325

 

Category results

Category IoU iIoU
flat 98.231 -
nature 91.2385 -
object 60.6957 -
sky 94.5225 -
construction 90.1287 -
human 75.89 54.3862
vehicle 92.1065 82.2246

 

Links

Download results as .csv file

Benchmark page