History

Miao Zheng 104429be00 [Docs] Replace markdownlint with mdformat for avoiding installing ruby (#1591 ) * [Docs] Replace markdownlint with mdformat for avoiding installing ruby * [Docs] Replace markdownlint with mdformat for avoiding installing ruby * lint		2022-05-20 18:29:44 +08:00
..
README.md	[Docs] Replace markdownlint with mdformat for avoiding installing ruby (#1591 )	2022-05-20 18:29:44 +08:00
fast_scnn_lr0.12_8x4_160k_cityscapes.py	[Fix] fix fast scnn (#606 )	2021-07-02 17:58:35 +08:00
fastscnn.yml	[Fix] Update correct `In Collection` in metafile of each configs. (#1239 )	2022-02-23 18:00:28 +08:00

README.md

Fast-SCNN

Fast-SCNN for Semantic Segmentation

Introduction

Official Repo

Code Snippet

Abstract

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.

Citation

@article{poudel2019fast,
  title={Fast-scnn: Fast semantic segmentation network},
  author={Poudel, Rudra PK and Liwicki, Stephan and Cipolla, Roberto},
  journal={arXiv preprint arXiv:1902.04502},
  year={2019}
}

Results and models

Cityscapes

Method	Backbone	Crop Size	Lr schd	Mem (GB)	Inf time (fps)	mIoU	mIoU(ms+flip)	config	download
FastSCNN	FastSCNN	512x1024	160000	3.3	56.45	70.96	72.65	config	model \| log