NL-LinkNet: Toward Lighter But More Accurate Road Extraction With Nonlocal Operations (2019-08-22T00:00:00.000000Z)

TL;DR

This letter proposes an efficient nonlocal LinkNet with nonlocal blocks (NLBs) that can grasp relations between global features that enables each spatial feature point to refer to all other contextual information and results in more accurate road segmentation.

Abstract

Road extraction from very high resolution (VHR) satellite images is one of the most important topics in the field of remote sensing. In this letter, we propose an efficient nonlocal LinkNet with nonlocal blocks (NLBs) that can grasp relations between global features. This enables each spatial feature point to refer to all other contextual information and results in more accurate road segmentation. In detail, our single model without any postprocessing like conditional random field (CRF) refinement performed better than any other published state-of-the-art ensemble model in the official DeepGlobe Challenge. Moreover, our nonlocal LinkNet (NL-LinkNet) beat the D-LinkNet, the winner of the DeepGlobe challenge (Demir et al., 2018), with 43% less parameters, less giga floating-point operations per seconds (GFLOPs), and shorter training convergence time. We also present empirical analyses on the proper usages of NLBs for the baseline model.

Authors

Yooseung Wang

1 papers

Junghoon Seo

2 papers

Taegyun Jeon

1 papers

TL;DR

Abstract

Authors

References35 items

Residual Non-local Attention Networks for Image Restoration

Attentional Road Safety Networks

Parameter-Free Spatial Attention Network for Person Re-Identification

Visual Spatial Attention Network for Relationship Detection

PSANet: Point-wise Spatial Attention Network for Scene Parsing

Non-Local Recurrent Network for Image Restoration

D-LinkNet: LinkNet with Pretrained Encoder and Dilated Convolution for High Resolution Satellite Imagery Road Extraction

Residual Inception Skip Network for Binary Segmentation

Road Detection with EOSResUNet and Post Vectorizing Algorithm

Stacked U-Nets with Multi-output for Road Extraction

DeepGlobe 2018: A Challenge to Parse the Earth through Satellite Images

Road Segmentation in SAR Satellite Images With Deep Fully Convolutional Neural Networks

Road Extraction by Deep Residual U-Net

Non-local Neural Networks

Automatic differentiation in PyTorch

Rethinking Atrous Convolution for Semantic Image Segmentation

LinkNet: Exploiting encoder representations for efficient semantic segmentation

Deformable Convolutional Networks

Road Structure Refined CNN for Road Extraction in Aerial Image

Convolutional Neural Networks for Large-Scale Remote-Sensing Image Classification

Understanding the Effective Receptive Field in Deep Convolutional Neural Networks

The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation

Fully convolutional networks for building and road extraction: Preliminary results

Deep Residual Learning for Image Recognition

Transferring Deep Convolutional Neural Networks for the Scene Classification of High-Resolution Remote Sensing Imagery

Adam: A Method for Stochastic Optimization

Fully convolutional networks for semantic segmentation

A non-local algorithm for image denoising

Road Extraction Using SVM and Image Segmentation

Detection of linear features in SAR images: application to road network extraction

A Computational Approach to Edge Detection

Use of the Hough transformation to detect lines and curves in pictures

Using Deep Convolutional Networks to Detect Roads in Aerial Images

Satrec Initiative Inc.

Machine Vision and Applications Automatic Extraction of Roads from Aerial Images Based on Scale Space and Snakes

Field of Study

Journal Information

Name

Page

Volume

Venue Information

Name

Type

URL

Alternate Names