Method Details


Details for method 'MSNET'

 

Method overview

name MSNET
challenge pixel-level semantic labeling
details
publication Anonymous
project page / code
used Cityscapes data fine annotations
used external data
runtime 0.2 s
GTX 1080
subsampling no
submission date October, 2017
previous submissions 1, 2

 

Average results

Metric Value
IoU Classes 72.1387
iIoU Classes 50.8114
IoU Categories 88.7618
iIoU Categories 77.2017

 

Class results

Class IoU iIoU
road 97.9697 -
sidewalk 82.2414 -
building 91.2658 -
wall 44.3287 -
fence 52.7758 -
pole 61.4472 -
traffic light 67.7265 -
traffic sign 72.8912 -
vegetation 92.5631 -
terrain 67.71 -
sky 94.3799 -
person 82.7391 67.78
rider 65.179 43.8317
car 94.4512 87.1865
truck 53.306 31.6151
bus 63.4199 43.7167
train 51.1903 33.402
motorcycle 62.86 39.5433
bicycle 72.1906 59.4159

 

Category results

Category IoU iIoU
flat 98.3729 -
nature 92.2187 -
object 67.9358 -
sky 94.3799 -
construction 91.722 -
human 83.1199 68.8909
vehicle 93.5834 85.5126

 

Links

Download results as .csv file

Benchmark page