Method Details


Details for method 'CLRCNet'

 

Method overview

name CLRCNet
challenge pixel-level semantic labeling
details A lightweight and real-time semantic segmentation method.
publication CLRCNet: Cascaded Low-Rank Convolutions for Semantic Segmentation in Real-time
project page / code
used Cityscapes data fine annotations
used external data
runtime 0.013 s
GPU
subsampling no
submission date June, 2018
previous submissions

 

Average results

Metric Value
IoU Classes 63.2943
iIoU Classes 35.9316
IoU Categories 84.4282
iIoU Categories 67.9667

 

Class results

Class IoU iIoU
road 97.1999 -
sidewalk 77.2915 -
building 88.0284 -
wall 35.9323 -
fence 40.1409 -
pole 50.1441 -
traffic light 55.0999 -
traffic sign 60.0478 -
vegetation 90.9384 -
terrain 66.7116 -
sky 93.4022 -
person 73.1222 51.8473
rider 49.8428 26.5969
car 90.8412 84.1609
truck 37.3517 13.6476
bus 51.4799 27.5139
train 45.1249 21.9504
motorcycle 41.9047 18.3464
bicycle 57.9865 43.3893

 

Category results

Category IoU iIoU
flat 97.9911 -
nature 90.5734 -
object 56.7124 -
sky 93.4022 -
construction 88.4998 -
human 74.0848 53.4458
vehicle 89.7337 82.4876

 

Links

Download results as .csv file

Benchmark page