Method Details

Details for method 'Fast-SCNN (Half-resolution)'


Method overview

name Fast-SCNN (Half-resolution)
challenge pixel-level semantic labeling
details The encoder-decoder framework is state-of-the-art for offline semantic image segmentation. Since the rise in autonomous systems, real-time computation is increasingly desirable. In this paper, we introduce fast segmentation convolutional neural network (Fast-SCNN), an above real-time semantic segmentation model on high resolution image data (1024x2048px) suited to efficient computation on embedded devices with low memory. Building on existing two-branch methods for fast segmentation, we introduce our `learning to downsample' module which computes low-level features for multiple resolution branches simultaneously. Our network combines spatial detail at high resolution with deep features extracted at lower resolution, yielding an accuracy of 68.0% mean intersection over union at 123.5 frames per second on Cityscapes. We also show that large scale pre-training is unnecessary. We thoroughly validate our metric in experiments with ImageNet pre-training and the coarse labeled data of Cityscapes. Finally, we show even faster computation with competitive results on subsampled inputs, without any network modifications.
publication Fast-SCNN: Fast Semantic Segmentation Network
Rudra P K Poudel, Stephan Liwicki, Roberto Cipolla
project page / code
used Cityscapes data fine annotations, coarse annotations
used external data
runtime 0.0035 s
Nvidia Titan Xp (Pascal)
subsampling 2
submission date November, 2018
previous submissions


Average results

Metric Value
IoU Classes 62.8271
iIoU Classes 31.9036
IoU Categories 80.5114
iIoU Categories 57.1292


Class results

Class IoU iIoU
road 97.393 -
sidewalk 77.7933 -
building 87.4057 -
wall 39.6781 -
fence 41.7532 -
pole 34.9656 -
traffic light 39.3722 -
traffic sign 50.4882 -
vegetation 88.506 -
terrain 63.3335 -
sky 92.6937 -
person 65.7443 36.7483
rider 46.3646 17.1671
car 91.0099 79.6037
truck 56.8599 16.755
bus 70.3143 27.0025
train 56.5166 29.919
motorcycle 40.8937 14.9045
bicycle 52.6286 33.1285


Category results

Category IoU iIoU
flat 97.6796 -
nature 88.0269 -
object 42.5333 -
sky 92.6937 -
construction 87.3266 -
human 65.893 37.6338
vehicle 89.4264 76.6246



Download results as .csv file

Benchmark page