Method Details


Details for method 'navi'

 

Method overview

name navi
challenge pixel-level semantic labeling
details mutil scale test
publication yuxb
project page / code
used Cityscapes data fine annotations
used external data ImageNet
runtime n/a
subsampling no
submission date December, 2018
previous submissions

 

Average results

Metric Value
IoU Classes 81.0235
iIoU Classes 60.0798
IoU Categories 91.1586
iIoU Categories 79.0667

 

Class results

Class IoU iIoU
road 98.4854 -
sidewalk 85.7039 -
building 93.4502 -
wall 61.7611 -
fence 60.0749 -
pole 68.1294 -
traffic light 75.3385 -
traffic sign 78.7106 -
vegetation 93.7319 -
terrain 73.5166 -
sky 95.6037 -
person 86.8395 66.9505
rider 69.6399 49.685
car 96.0848 91.4706
truck 78.3067 47.3838
bus 92.6571 52.2124
train 87.0375 60.7049
motorcycle 68.7811 47.4774
bicycle 75.5944 64.7541

 

Category results

Category IoU iIoU
flat 98.7543 -
nature 93.4718 -
object 73.9777 -
sky 95.6037 -
construction 93.7259 -
human 86.9657 68.2082
vehicle 95.6108 89.9251

 

Links

Download results as .csv file

Benchmark page