Method Details


Details for method 'EFRNet-13'

 

Method overview

name EFRNet-13
challenge pixel-level semantic labeling
details
publication Anonymous
project page / code
used Cityscapes data fine annotations
used external data
runtime 0.0146 s
Intel i7 7700, NVIDIA TitanX 12G
subsampling no
submission date August, 2019
previous submissions

 

Average results

Metric Value
IoU Classes 72.8236
iIoU Classes 43.6333
IoU Categories 86.9955
iIoU Categories 70.0678

 

Class results

Class IoU iIoU
road 98.1768 -
sidewalk 83.0216 -
building 90.8049 -
wall 56.5252 -
fence 51.3743 -
pole 54.8893 -
traffic light 60.4275 -
traffic sign 66.382 -
vegetation 92.0904 -
terrain 69.5998 -
sky 94.6693 -
person 78.615 56.2382
rider 59.4031 31.999
car 93.548 85.3629
truck 62.7822 23.3536
bus 79.1734 38.6051
train 72.715 36.6929
motorcycle 53.5877 29.7738
bicycle 65.8633 47.0408

 

Category results

Category IoU iIoU
flat 98.3298 -
nature 91.7865 -
object 61.5177 -
sky 94.6693 -
construction 91.037 -
human 78.917 57.4738
vehicle 92.7115 82.6619

 

Links

Download results as .csv file

Benchmark page