Method Details


Details for method 'SAIT SeeThroughNet'

 

Method overview

name SAIT SeeThroughNet
challenge pixel-level semantic labeling
details
publication Anonymous
project page / code
used Cityscapes data fine annotations, coarse annotations
used external data ImageNet, Mapillary
runtime n/a
subsampling no
submission date November, 2021
previous submissions 1, 2

 

Average results

Metric Value
IoU Classes 86.1152
iIoU Classes 71.865
IoU Categories 93.1897
iIoU Categories 85.6068

 

Class results

Class IoU iIoU
road 98.9444 -
sidewalk 89.0935 -
building 95.0449 -
wall 73.755 -
fence 71.3477 -
pole 76.0951 -
traffic light 82.4122 -
traffic sign 85.1273 -
vegetation 94.52 -
terrain 75.2644 -
sky 96.2455 -
person 90.0618 78.8276
rider 79.6302 64.8131
car 97.0403 92.3268
truck 84.2469 62.5732
bus 95.0706 72.3899
train 92.6783 67.5829
motorcycle 77.8477 63.1125
bicycle 81.7635 73.2938

 

Category results

Category IoU iIoU
flat 98.9098 -
nature 94.2988 -
object 81.0078 -
sky 96.2455 -
construction 95.0041 -
human 90.1016 79.8283
vehicle 96.7601 91.3853

 

Links

Download results as .csv file

Benchmark page