Method Details


Details for method 'GALD-net'

 

Method overview

name GALD-net
challenge pixel-level semantic labeling
details We propose Global Aggregation then Local Distribution (GALD) scheme to distribute global information to each position adaptively according the local information surrounding the position.
publication Global Aggregation then Local Distribution for Scene Understanding
Xiangtai Li, Maoke Yang, Ansheng You, Kuiyuan Yang
project page / code https://github.com/lxtGH/GALD_net
used Cityscapes data fine annotations, coarse annotations
used external data Mapillary
runtime n/a
subsampling no
submission date February, 2019
previous submissions

 

Average results

Metric Value
IoU Classes 83.1276
iIoU Classes 63.4849
IoU Categories 92.2452
iIoU Categories 81.4271

 

Class results

Class IoU iIoU
road 98.7948 -
sidewalk 87.5802 -
building 94.1835 -
wall 64.5675 -
fence 66.5406 -
pole 72.8465 -
traffic light 79.0412 -
traffic sign 82.217 -
vegetation 94.1592 -
terrain 73.0619 -
sky 96.0223 -
person 88.3038 72.8503
rider 75.2141 55.7875
car 96.4999 90.9287
truck 79.3227 51.198
bus 89.7172 58.5384
train 87.5446 58.7452
motorcycle 73.8065 52.5305
bicycle 80.0008 67.3003

 

Category results

Category IoU iIoU
flat 98.7853 -
nature 93.8186 -
object 78.0226 -
sky 96.0223 -
construction 94.4141 -
human 88.5138 73.7503
vehicle 96.1398 89.1039

 

Links

Download results as .csv file

Benchmark page