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 September, 2016
previous submissions 1

 

Average results

Metric Value
IoU Classes 76.1309
iIoU Classes 51.2074
IoU Categories 89.645
iIoU Categories 75.0651

 

Class results

Class IoU iIoU
road 98.4331 -
sidewalk 84.8493 -
building 92.3709 -
wall 54.2907 -
fence 54.2678 -
pole 62.7729 -
traffic light 70.1917 -
traffic sign 75.9268 -
vegetation 93.1913 -
terrain 70.9047 -
sky 94.7463 -
person 84.1469 62.629
rider 66.471 42.503
car 95.2862 88.2116
truck 68.2743 34.1422
bus 78.04 46.4452
train 63.9482 40.8469
motorcycle 64.8195 39.0137
bicycle 73.5554 55.8674

 

Category results

Category IoU iIoU
flat 98.5497 -
nature 92.826 -
object 69.5285 -
sky 94.7463 -
construction 92.7728 -
human 84.5057 64.0596
vehicle 94.5862 86.0707

 

Links

Download results as .csv file

Benchmark page