Method Details


Details for method 'MiniNet-v2-pretrained'

 

Method overview

name MiniNet-v2-pretrained
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
subsampling 2
submission date February, 2019
previous submissions

 

Average results

Metric Value
IoU Classes 68.0152
iIoU Classes 39.7194
IoU Categories 86.2011
iIoU Categories 68.3539

 

Class results

Class IoU iIoU
road 97.9671 -
sidewalk 81.5904 -
building 89.9299 -
wall 43.4969 -
fence 44.3058 -
pole 54.7678 -
traffic light 57.9584 -
traffic sign 64.1489 -
vegetation 91.6813 -
terrain 68.7597 -
sky 94.5139 -
person 75.9027 52.117
rider 54.2737 29.6539
car 93.1938 85.3966
truck 49.1187 20.0851
bus 67.6779 33.4999
train 57.0396 30.7938
motorcycle 44.95 23.3874
bicycle 61.0118 42.8218

 

Category results

Category IoU iIoU
flat 98.254 -
nature 91.2799 -
object 60.8866 -
sky 94.5139 -
construction 90.176 -
human 76.1854 53.7771
vehicle 92.112 82.9306

 

Links

Download results as .csv file

Benchmark page