Method Details


Details for method 'ALV303'

 

Method overview

name ALV303
challenge pixel-level semantic labeling
details
publication Anonymous
project page / code
used Cityscapes data fine annotations
used external data ImageNet
runtime 0.2 s
GTX 1080ti
subsampling no
submission date July, 2018
previous submissions

 

Average results

Metric Value
IoU Classes 72.2458
iIoU Classes 52.0009
IoU Categories 89.7929
iIoU Categories 79.2308

 

Class results

Class IoU iIoU
road 98.341 -
sidewalk 84.2211 -
building 91.9838 -
wall 42.0616 -
fence 51.7817 -
pole 65.9323 -
traffic light 73.6756 -
traffic sign 77.4984 -
vegetation 92.8821 -
terrain 71.559 -
sky 94.9591 -
person 83.0072 67.6903
rider 61.6112 47.3534
car 94.5817 90.1826
truck 46.8268 31.1221
bus 60.0485 46.0979
train 54.3394 33.9626
motorcycle 55.524 40.0776
bicycle 71.8356 59.5208

 

Category results

Category IoU iIoU
flat 98.557 -
nature 92.5388 -
object 72.242 -
sky 94.9591 -
construction 92.1136 -
human 84.3823 69.8768
vehicle 93.7573 88.5848

 

Links

Download results as .csv file

Benchmark page