Method Details


Details for method 'RFMobileNetV2Plus'

 

Method overview

name RFMobileNetV2Plus
challenge pixel-level semantic labeling
details Receptive Filed MobileNetV2Plus for Semantic Segmentation
publication Huijun Liu
project page / code
used Cityscapes data fine annotations
used external data ImageNet
runtime n/a
subsampling no
submission date March, 2018
previous submissions

 

Average results

Metric Value
IoU Classes 70.6749
iIoU Classes 49.4492
IoU Categories 88.2688
iIoU Categories 75.8143

 

Class results

Class IoU iIoU
road 97.7627 -
sidewalk 80.6175 -
building 90.8641 -
wall 44.2851 -
fence 47.0401 -
pole 59.5487 -
traffic light 65.6004 -
traffic sign 72.2945 -
vegetation 92.514 -
terrain 68.5927 -
sky 94.1159 -
person 81.5643 64.2324
rider 60.6195 44.3144
car 94.3278 87.299
truck 51.2618 31.3601
bus 61.2393 39.6292
train 57.068 34.0697
motorcycle 54.0468 37.5761
bicycle 69.4594 57.1124

 

Category results

Category IoU iIoU
flat 98.3544 -
nature 92.2254 -
object 66.6388 -
sky 94.1159 -
construction 91.4593 -
human 82.0359 66.3245
vehicle 93.0522 85.3041

 

Links

Download results as .csv file

Benchmark page