Fast-SCNN: Fast Semantic Segmentation Network

12 Feb 2019  ·  Rudra P. K. Poudel, Stephan Liwicki, Roberto Cipolla ·

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.

PDF Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Semantic Segmentation Cityscapes test Fast-SCNN Mean IoU (class) 68% # 83
Semantic Segmentation Cityscapes val Fast-SCNN + Coarse + ImageNet mIoU 69.19 # 83
Semantic Segmentation DADA-seg Fast-SCNN mIoU 26.32 # 13
Semantic Segmentation DensePASS Fast-SCNN mIoU 24.6% # 34
Semantic Segmentation EventScape Fast-SCNN mIoU 44.27 # 9
Thermal Image Segmentation PST900 Fast-SCNN mIoU 47.2 # 20
Semantic Segmentation SynPASS Fast-SCNN mIoU 21.30% # 6

Methods


No methods listed for this paper. Add relevant methods here