Method Details


Details for method 'Bilateral_attention_semantic'

 

Method overview

name Bilateral_attention_semantic
challenge pixel-level semantic labeling
details we use bilateral attention mechanism for semantic segmentation
publication Anonymous
project page / code
used Cityscapes data fine annotations
used external data
runtime 0.0141 s
Nvidia Tesla V100
subsampling no
submission date September, 2020
previous submissions 1, 2

 

Average results

Metric Value
IoU Classes 76.4899
iIoU Classes 55.9409
IoU Categories 90.3647
iIoU Categories 79.4577

 

Class results

Class IoU iIoU
road 98.4219 -
sidewalk 84.9274 -
building 92.5435 -
wall 48.0638 -
fence 55.4155 -
pole 65.6648 -
traffic light 73.8527 -
traffic sign 77.2465 -
vegetation 93.257 -
terrain 71.704 -
sky 94.9367 -
person 85.5992 69.7131
rider 68.903 47.2872
car 95.4002 89.6569
truck 62.7195 34.7948
bus 77.5014 50.6313
train 70.9734 47.1676
motorcycle 61.5068 44.0372
bicycle 74.6705 64.239

 

Category results

Category IoU iIoU
flat 98.6403 -
nature 92.9416 -
object 71.969 -
sky 94.9367 -
construction 92.9042 -
human 86.1183 70.9903
vehicle 95.0429 87.9251

 

Links

Download results as .csv file

Benchmark page