Method Details


Details for method 'FAN_NV_RVC'

 

Method overview

name FAN_NV_RVC
challenge pixel-level semantic labeling
details Hybrid-Base + Segformer
publication Anonymous
project page / code
used Cityscapes data fine annotations
used external data ADE20k, COCO, MVD, ScanNet, VIPER, WildDash2, MSeg
runtime n/a
subsampling no
submission date September, 2022
previous submissions

 

Average results

Metric Value
IoU Classes 81.9729
iIoU Classes 59.0079
IoU Categories 91.0244
iIoU Categories 78.2816

 

Class results

Class IoU iIoU
road 98.4495 -
sidewalk 84.7723 -
building 93.818 -
wall 65.059 -
fence 66.8069 -
pole 67.272 -
traffic light 74.5346 -
traffic sign 78.0021 -
vegetation 94.0302 -
terrain 72.2493 -
sky 96.1283 -
person 86.7671 67.2082
rider 71.2794 46.238
car 96.2769 90.607
truck 81.054 44.2282
bus 93.2888 58.4307
train 88.6484 55.1428
motorcycle 72.1559 49.3133
bicycle 76.8917 60.8948

 

Category results

Category IoU iIoU
flat 97.908 -
nature 93.5617 -
object 73.0487 -
sky 96.1283 -
construction 93.6421 -
human 87.1236 68.078
vehicle 95.7582 88.4852

 

Links

Download results as .csv file

Benchmark page