Skip to content

Latest commit

 

History

History
63 lines (44 loc) · 12.4 KB

README.md

File metadata and controls

63 lines (44 loc) · 12.4 KB

FastFCN

FastFCN: Rethinking Dilated Convolution in the Backbone for Semantic Segmentation

Introduction

Official Repo

Code Snippet

Abstract

Modern approaches for semantic segmentation usually employ dilated convolutions in the backbone to extract high-resolution feature maps, which brings heavy computation complexity and memory footprint. To replace the time and memory consuming dilated convolutions, we propose a novel joint upsampling module named Joint Pyramid Upsampling (JPU) by formulating the task of extracting high-resolution feature maps into a joint upsampling problem. With the proposed JPU, our method reduces the computation complexity by more than three times without performance loss. Experiments show that JPU is superior to other upsampling modules, which can be plugged into many existing approaches to reduce computation complexity and improve performance. By replacing dilated convolutions with the proposed JPU module, our method achieves the state-of-the-art performance in Pascal Context dataset (mIoU of 53.13%) and ADE20K dataset (final score of 0.5584) while running 3 times faster.

Results and models

Cityscapes

Method Backbone Crop Size Lr schd Mem (GB) Inf time (fps) Device mIoU mIoU(ms+flip) config download
FastFCN + DeepLabV3 R-50-D32 512x1024 80000 5.67 2.64 V100 79.12 80.58 config model | log
FastFCN + DeepLabV3 R-50-D32 (4x4) 512x1024 80000 9.79 - V100 79.52 80.91 config model | log
FastFCN + PSPNet R-50-D32 512x1024 80000 5.67 4.40 V100 79.26 80.86 config model | log
FastFCN + PSPNet R-50-D32 (4x4) 512x1024 80000 9.94 - V100 78.76 80.03 config model | log
FastFCN + EncNet R-50-D32 512x1024 80000 8.15 4.77 V100 77.97 79.92 config model | log
FastFCN + EncNet R-50-D32 (4x4) 512x1024 80000 15.45 - V100 78.6 80.25 config model | log

ADE20K

Method Backbone Crop Size Lr schd Mem (GB) Inf time (fps) Device mIoU mIoU(ms+flip) config download
FastFCN + DeepLabV3 R-50-D32 512x1024 80000 8.46 12.06 V100 41.88 42.91 config model | log
FastFCN + DeepLabV3 R-50-D32 512x1024 160000 - - V100 43.58 44.92 config model | log
FastFCN + PSPNet R-50-D32 512x1024 80000 8.02 19.21 V100 41.40 42.12 config model | log
FastFCN + PSPNet R-50-D32 512x1024 160000 - - V100 42.63 43.71 config model | log
FastFCN + EncNet R-50-D32 512x1024 80000 9.67 17.23 V100 40.88 42.36 config model | log
FastFCN + EncNet R-50-D32 512x1024 160000 - - V100 42.50 44.21 config model | log

Note:

Citation

@article{wu2019fastfcn,
title={Fastfcn: Rethinking dilated convolution in the backbone for semantic segmentation},
author={Wu, Huikai and Zhang, Junge and Huang, Kaiqi and Liang, Kongming and Yu, Yizhou},
journal={arXiv preprint arXiv:1903.11816},
year={2019}
}