Method Details


Details for method 'TuSimple'

 

Method overview

name TuSimple
challenge pixel-level semantic labeling
details
publication Anonymous
project page / code
used Cityscapes data fine annotations
used external data ImageNet
runtime n/a
subsampling no
submission date August, 2016
previous submissions

 

Average results

Metric Value
IoU Classes 75.9656
iIoU Classes 50.8658
IoU Categories 89.4689
iIoU Categories 74.8164

 

Class results

Class IoU iIoU
road 98.4128 -
sidewalk 84.8605 -
building 92.3832 -
wall 54.4796 -
fence 53.9389 -
pole 61.4125 -
traffic light 69.734 -
traffic sign 75.2401 -
vegetation 93.0222 -
terrain 70.4788 -
sky 94.7546 -
person 84.2089 62.7195
rider 66.5119 42.3775
car 95.2898 87.8889
truck 69.8664 34.6891
bus 77.1414 45.9209
train 63.0298 37.6536
motorcycle 65.3116 39.4897
bicycle 73.269 56.1874

 

Category results

Category IoU iIoU
flat 98.5274 -
nature 92.6789 -
object 68.5798 -
sky 94.7546 -
construction 92.7274 -
human 84.4449 63.9068
vehicle 94.5694 85.7259

 

Links

Download results as .csv file

Benchmark page