Method Details


Details for method 'MeiTuan-BaseModel'

 

Method overview

name MeiTuan-BaseModel
challenge pixel-level semantic labeling
details previously listed as "MT-base_cv"
publication Anonymous
project page / code
used Cityscapes data fine annotations, coarse annotations
used external data ImageNet
runtime n/a
subsampling no
submission date July, 2022
previous submissions

 

Average results

Metric Value
IoU Classes 86.367
iIoU Classes 72.9695
IoU Categories 93.3546
iIoU Categories 85.3598

 

Class results

Class IoU iIoU
road 98.9354 -
sidewalk 89.4852 -
building 95.0214 -
wall 72.9002 -
fence 72.3598 -
pole 76.6546 -
traffic light 82.1966 -
traffic sign 85.1959 -
vegetation 94.6285 -
terrain 75.5691 -
sky 96.3544 -
person 90.5926 78.8005
rider 80.7766 64.9547
car 96.8319 91.8256
truck 84.2275 61.9984
bus 95.4048 74.8001
train 92.074 71.9883
motorcycle 79.1841 65.4385
bicycle 82.5798 73.95

 

Category results

Category IoU iIoU
flat 98.9268 -
nature 94.4689 -
object 81.3625 -
sky 96.3544 -
construction 95.172 -
human 90.5964 79.6015
vehicle 96.6014 91.118

 

Links

Download results as .csv file

Benchmark page