Method Details


Details for method 'PSPNet'

 

Method overview

name PSPNet
challenge pixel-level semantic labeling
details Using training set only (2975 images) without adding validation set or coarse set, no post-processsing like CRF.
publication Anonymous
project page / code
used Cityscapes data fine annotations
used external data ImageNet
runtime n/a
subsampling no
submission date August, 2016
previous submissions 1, 2, 3, 4

 

Average results

Metric Value
IoU Classes 76.0191
iIoU Classes 52.3949
IoU Categories 89.8889
iIoU Categories 76.9949

 

Class results

Class IoU iIoU
road 98.4688 -
sidewalk 85.1647 -
building 92.6668 -
wall 49.7215 -
fence 58.1827 -
pole 61.9611 -
traffic light 72.5287 -
traffic sign 76.7125 -
vegetation 93.0693 -
terrain 71.9124 -
sky 95.1672 -
person 85.3427 65.4822
rider 68.199 45.0622
car 95.3961 89.7014
truck 56.6668 29.9054
bus 72.7411 48.309
train 68.8423 37.8376
motorcycle 66.4095 43.1447
bicycle 75.2092 59.7164

 

Category results

Category IoU iIoU
flat 98.6405 -
nature 92.7546 -
object 69.3322 -
sky 95.1672 -
construction 92.7693 -
human 85.7883 66.5106
vehicle 94.7703 87.4792

 

Links

Download results as .csv file

Benchmark page