Method Details


Details for method 'MNet_MPRG'

 

Method overview

name MNet_MPRG
challenge pixel-level semantic labeling
details without val dataset, external dataset (e.g. image net) and post-processing
publication Chubu University, MPRG
project page / code
used Cityscapes data fine annotations
used external data
runtime 0.6 s
GPU
subsampling no
submission date November, 2016
previous submissions

 

Average results

Metric Value
IoU Classes 71.9465
iIoU Classes 46.5843
IoU Categories 89.2759
iIoU Categories 77.8834

 

Class results

Class IoU iIoU
road 98.1494 -
sidewalk 82.9311 -
building 91.818 -
wall 43.6378 -
fence 50.5122 -
pole 64.3176 -
traffic light 71.4036 -
traffic sign 74.6109 -
vegetation 92.6852 -
terrain 70.2801 -
sky 94.734 -
person 82.4462 66.9241
rider 60.8808 40.6319
car 94.128 89.1923
truck 50.9138 22.4619
bus 62.4706 32.8811
train 57.1748 30.0706
motorcycle 53.8392 33.3182
bicycle 70.0496 57.1939

 

Category results

Category IoU iIoU
flat 98.5324 -
nature 92.4052 -
object 70.4231 -
sky 94.734 -
construction 91.9254 -
human 83.3112 68.6196
vehicle 93.6002 87.1471

 

Links

Download results as .csv file

Benchmark page