Method Details


Details for method 'Hyundai Mobis AD Lab'

 

Method overview

name Hyundai Mobis AD Lab
challenge pixel-level semantic labeling
details
publication Hyundai Mobis AD Lab, DL-DB Group, AA (Automated Annotator) Team
project page / code
used Cityscapes data fine annotations, coarse annotations
used external data ImageNet, Mapillary
runtime n/a
subsampling no
submission date August, 2019
previous submissions

 

Average results

Metric Value
IoU Classes 83.8384
iIoU Classes 65.0297
IoU Categories 92.3943
iIoU Categories 82.3911

 

Class results

Class IoU iIoU
road 98.8709 -
sidewalk 88.3606 -
building 94.317 -
wall 65.2046 -
fence 65.9264 -
pole 72.8185 -
traffic light 79.5029 -
traffic sign 82.9659 -
vegetation 94.2762 -
terrain 74.3217 -
sky 96.1378 -
person 88.614 73.6769
rider 75.8681 55.3217
car 96.6408 92.1782
truck 79.3213 52.2287
bus 93.8375 61.3387
train 91.5037 63.8842
motorcycle 74.7612 54.6461
bicycle 79.6809 66.9631

 

Category results

Category IoU iIoU
flat 98.8418 -
nature 94.0253 -
object 78.2646 -
sky 96.1378 -
construction 94.5254 -
human 88.6688 74.3554
vehicle 96.2967 90.4268

 

Links

Download results as .csv file

Benchmark page