Method Details


Details for method 'SegModel'

 

Method overview

name SegModel
challenge pixel-level semantic labeling
details
publication Anonymous
project page / code
used Cityscapes data fine annotations, coarse annotations
used external data ImageNet, MSCOCO
runtime n/a
subsampling no
submission date November, 2016
previous submissions 1, 2

 

Average results

Metric Value
IoU Classes 79.2067
iIoU Classes 56.4295
IoU Categories 90.3522
iIoU Categories 77.0434

 

Class results

Class IoU iIoU
road 98.5906 -
sidewalk 86.2351 -
building 93.0094 -
wall 53.7243 -
fence 60.3695 -
pole 64.1637 -
traffic light 73.5069 -
traffic sign 78.5456 -
vegetation 93.4307 -
terrain 72.171 -
sky 95.5415 -
person 85.3177 65.2085
rider 68.5686 46.7258
car 95.8423 90.1237
truck 77.9018 41.5115
bus 86.96 55.053
train 77.9577 48.3188
motorcycle 68.0404 45.3316
bicycle 75.0501 59.1635

 

Category results

Category IoU iIoU
flat 98.6509 -
nature 93.1595 -
object 71.175 -
sky 95.5415 -
construction 93.4518 -
human 85.3277 66.1701
vehicle 95.1587 87.9167

 

Links

Download results as .csv file

Benchmark page