Method Details


Details for method 'SAIT'

 

Method overview

name SAIT
challenge pixel-level semantic labeling
details only train data (2,975), full resolution training
publication Anonymous
project page / code
used Cityscapes data fine annotations
used external data ImageNet
runtime 1 s
GPU TITAN X
subsampling no
submission date April, 2016
previous submissions 1

 

Average results

Metric Value
IoU Classes 67.6896
iIoU Classes 40.1272
IoU Categories 86.3304
iIoU Categories 68.1926

 

Class results

Class IoU iIoU
road 97.8639 -
sidewalk 80.6982 -
building 89.8794 -
wall 40.2122 -
fence 40.0262 -
pole 52.8634 -
traffic light 59.5246 -
traffic sign 66.8614 -
vegetation 91.2207 -
terrain 68.876 -
sky 94.2774 -
person 77.7787 54.6931
rider 57.6591 30.7798
car 93.1247 82.3951
truck 45.1175 20.8157
bus 56.3322 28.6728
train 52.6442 26.3903
motorcycle 54.7768 30.6116
bicycle 66.3651 46.6589

 

Category results

Category IoU iIoU
flat 98.2938 -
nature 90.9371 -
object 60.3807 -
sky 94.2774 -
construction 90.0559 -
human 78.376 56.1228
vehicle 91.9919 80.2623

 

Links

Download results as .csv file

Benchmark page