Regularization of Building Boundaries in Satellite Images Using Adversarial and Regularized Losses

Zorzi, Stefano and Fraundorfer, Friedrich (2019) Regularization of Building Boundaries in Satellite Images Using Adversarial and Regularized Losses. In: 2019 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), pp. 5140-5143. IEEE. IGARSS 2019, 2019-07-28 - 2019-08-02, Yokohama, Japan. doi: 10.1109/IGARSS.2019.8900337. ISBN 978-1-5386-9154-0.

PDF - Only accessible within DLR
2MB

Official URL: https://ieeexplore.ieee.org/abstract/document/8900337

Abstract

In this paper we present a method for building boundary re-finement and regularization in satellite images using a fullyconvolutional neural network trained with a combination ofadversarial and regularized losses. Compared to a pure MaskR-CNN model, the overall algorithm can achieve equivalentperformance in terms of accuracy and completeness. How-ever, unlike Mask R-CNN that produces irregular footprints,our framework generates regularized and visually pleasingbuilding boundaries which are beneficial in many applica-tions.

Item URL in elib:

https://elib.dlr.de/132354/

Document Type:

Conference or Workshop Item (Speech)

Title:

Regularization of Building Boundaries in Satellite Images Using Adversarial and Regularized Losses

Authors:

Authors	Institution or Email of Authors	Author's ORCID iD	ORCID Put Code
Zorzi, Stefano	zorzi (at) icg.tugraz.at	UNSPECIFIED	UNSPECIFIED
Fraundorfer, Friedrich	friedrich.fraundorfer (at) dlr.de	https://orcid.org/0000-0002-5805-8892	UNSPECIFIED

Date:

July 2019

Journal or Publication Title:

2019 IEEE International Geoscience and Remote Sensing Symposium (IGARSS)

Refereed publication:

Yes

Open Access:

Gold Open Access:

In SCOPUS:

In ISI Web of Science:

DOI:

10.1109/IGARSS.2019.8900337

Page Range:

pp. 5140-5143

Publisher:

IEEE

ISBN:

978-1-5386-9154-0

Status:

Published

Keywords:

Generative adversarial networks, build-ing segmentation, boundary refinement, satellite images

Event Title:

IGARSS 2019

Event Location:

Yokohama, Japan

Event Type:

international Conference

Event Start Date:

28 July 2019

Event End Date:

2 August 2019

Organizer:

IEEE Geoscience and Remote Sensing Society (IGARSS)

HGF - Research field:

Aeronautics, Space and Transport

HGF - Program:

Transport

HGF - Program Themes:

Road Transport

DLR - Research area:

Transport

DLR - Program:

V ST Straßenverkehr

DLR - Research theme (Project):

V - NGC KoFiF (old)

Location:

Oberpfaffenhofen

Institutes and Institutions:

Remote Sensing Technology Institute > Photogrammetry and Image Analysis

Deposited By:

Reinartz, Prof. Dr.. Peter

Deposited On:

06 Dec 2019 12:41

Last Modified:

24 Apr 2024 20:35

Repository Staff Only: item control page