DisasterGAN: Generative Adversarial Networks for Remote Sensing Disaster Image Generation

Rapid progress on disaster detection and assessment has been achieved with the development of deep-learning techniques and the wide applications of remote sensing images. However, it is still a great challenge to train an accurate and robust disaster detection network due to the class imbalance of e...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Xue Rui, Yang Cao, Xin Yuan, Yu Kang, Weiguo Song
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	GAN image generation data augmentation remote sensing disaster image Science Q
Acceso en línea:	https://doaj.org/article/c0619a0b74ed40bdb748e1496cd3b225
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:c0619a0b74ed40bdb748e1496cd3b225
record_format	dspace
spelling	oai:doaj.org-article:c0619a0b74ed40bdb748e1496cd3b2252021-11-11T18:52:56ZDisasterGAN: Generative Adversarial Networks for Remote Sensing Disaster Image Generation10.3390/rs132142842072-4292https://doaj.org/article/c0619a0b74ed40bdb748e1496cd3b2252021-10-01T00:00:00Zhttps://www.mdpi.com/2072-4292/13/21/4284https://doaj.org/toc/2072-4292Rapid progress on disaster detection and assessment has been achieved with the development of deep-learning techniques and the wide applications of remote sensing images. However, it is still a great challenge to train an accurate and robust disaster detection network due to the class imbalance of existing data sets and the lack of training data. This paper aims at synthesizing disaster remote sensing images with multiple disaster types and different building damage with generative adversarial networks (GANs), making up for the shortcomings of the existing data sets. However, existing models are inefficient in multi-disaster image translation due to the diversity of disaster and inevitably change building-irrelevant regions caused by directly operating on the whole image. Thus, we propose two models: disaster translation GAN can generate disaster images for multiple disaster types using only a single model, which uses an attribute to represent disaster types and a reconstruction process to further ensure the effect of the generator; damaged building generation GAN is a mask-guided image generation model, which can only alter the attribute-specific region while keeping the attribute-irrelevant region unchanged. Qualitative and quantitative experiments demonstrate the validity of the proposed methods. Further experimental results on the damaged building assessment model show the effectiveness of the proposed models and the superiority compared with other data augmentation methods.Xue RuiYang CaoXin YuanYu KangWeiguo SongMDPI AGarticleGANimage generationdata augmentationremote sensing disaster imageScienceQENRemote Sensing, Vol 13, Iss 4284, p 4284 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	GAN image generation data augmentation remote sensing disaster image Science Q
spellingShingle	GAN image generation data augmentation remote sensing disaster image Science Q Xue Rui Yang Cao Xin Yuan Yu Kang Weiguo Song DisasterGAN: Generative Adversarial Networks for Remote Sensing Disaster Image Generation
description	Rapid progress on disaster detection and assessment has been achieved with the development of deep-learning techniques and the wide applications of remote sensing images. However, it is still a great challenge to train an accurate and robust disaster detection network due to the class imbalance of existing data sets and the lack of training data. This paper aims at synthesizing disaster remote sensing images with multiple disaster types and different building damage with generative adversarial networks (GANs), making up for the shortcomings of the existing data sets. However, existing models are inefficient in multi-disaster image translation due to the diversity of disaster and inevitably change building-irrelevant regions caused by directly operating on the whole image. Thus, we propose two models: disaster translation GAN can generate disaster images for multiple disaster types using only a single model, which uses an attribute to represent disaster types and a reconstruction process to further ensure the effect of the generator; damaged building generation GAN is a mask-guided image generation model, which can only alter the attribute-specific region while keeping the attribute-irrelevant region unchanged. Qualitative and quantitative experiments demonstrate the validity of the proposed methods. Further experimental results on the damaged building assessment model show the effectiveness of the proposed models and the superiority compared with other data augmentation methods.
format	article
author	Xue Rui Yang Cao Xin Yuan Yu Kang Weiguo Song
author_facet	Xue Rui Yang Cao Xin Yuan Yu Kang Weiguo Song
author_sort	Xue Rui
title	DisasterGAN: Generative Adversarial Networks for Remote Sensing Disaster Image Generation
title_short	DisasterGAN: Generative Adversarial Networks for Remote Sensing Disaster Image Generation
title_full	DisasterGAN: Generative Adversarial Networks for Remote Sensing Disaster Image Generation
title_fullStr	DisasterGAN: Generative Adversarial Networks for Remote Sensing Disaster Image Generation
title_full_unstemmed	DisasterGAN: Generative Adversarial Networks for Remote Sensing Disaster Image Generation
title_sort	disastergan: generative adversarial networks for remote sensing disaster image generation
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/c0619a0b74ed40bdb748e1496cd3b225
work_keys_str_mv	AT xuerui disastergangenerativeadversarialnetworksforremotesensingdisasterimagegeneration AT yangcao disastergangenerativeadversarialnetworksforremotesensingdisasterimagegeneration AT xinyuan disastergangenerativeadversarialnetworksforremotesensingdisasterimagegeneration AT yukang disastergangenerativeadversarialnetworksforremotesensingdisasterimagegeneration AT weiguosong disastergangenerativeadversarialnetworksforremotesensingdisasterimagegeneration
_version_	1718431725659357184

DisasterGAN: Generative Adversarial Networks for Remote Sensing Disaster Image Generation

Ejemplares similares