Identifying Disaster Regions in Images Through Attention Shifting With a Retarget Network

Thanyawet, Narongthat; Ratsamee, Photchara; Uranishi, Yuki; Kobayashi, Masato; Takemura, Haruo

doi:https://doi.org/10.1109/ACCESS.2024.3443130

Identifying Disaster Regions in Images Through Attention Shifting With a Retarget Network

Thanyawet, Narongthat; Ratsamee, Photchara; Uranishi, Yuki et al.

IEEE Access, 2024, 12, 143754-143766

アクセス数:787件（2025-08-15 19:12 集計）

固定URL: https://hdl.handle.net/11094/99687

閲覧可能ファイル

ファイル	フォーマット	利用条件	サイズ	閲覧回数	利用開始日	説明	information
IEEEAccess_12_143754	pdf	なし	3.11 MB	58	2025.01.20

Creative Commons : 表示 - 非営利 - 改変禁止

論文情報

ファイル出力 EndNote Basic出力 Mendeley出力

タイトル	Identifying Disaster Regions in Images Through Attention Shifting With a Retarget Network
著者	著者名 Thanyawet, Narongthat
	NRID 1000050772448 1000050772448 ORCID 0000-0002-3081-2232 0000-0002-3081-2232 researchmap permalink photchara photchara Scopus Author ID 55193483300 55193483300 著者名 Ratsamee, Photchara
	NRID 1000000533738 1000000533738 researchmap permalink yukiuranishi yukiuranishi Scopus Author ID 35321680700 35321680700 著者名 Uranishi, Yuki
	NRID 1000010979868 1000010979868 ORCID 0000-0001-9703-2858 0000-0001-9703-2858 researchmap permalink kobayashi.masato kobayashi.masato 著者名 Kobayashi, Masato
	NRID 1000060263430 1000060263430 ORCID 0000-0002-4645-6439 0000-0002-4645-6439 researchmap permalink read0089576 read0089576 Scopus Author ID 7201655004 7201655004 著者名 Takemura, Haruo
キーワード等	anchor box
	disaster detection
	image captioning
	Retarget network
	vision transformer
内容	Thanyawet N., Ratsamee P., Uranishi Y., et al. Identifying Disaster Regions in Images Through Attention Shifting With a Retarget Network. IEEE Access 12, 143754 (2024); https://doi.org/10.1109/ACCESS.2024.3443130.
抄録	Disasters disrupt lives and necessitate quick location of affected areas for rescue efforts. The application of computer vision has enhanced disaster detection, such as landslides and floods; however traditional computer vision methods often overlook smaller, critical details in favor of prominent objects. This research introduces the Retarget Network (RetNet), a novel framework aimed at improving image captioning techniques to identify and prioritize these less evident, yet crucial, objects in disaster scenarios, enhancing scene recognition and aiding in more effective disaster response. By masking images, we direct the model's focus towards additional significant areas within the image. RetNet employs anchor boxes to refine the targeting of specific areas and, optimize their center positions, heights, and widths. Additionally, RetNet determines which anchors to mask prior to captioning, facilitating the identification of challenging objects such as boulders, soil, and water, which are indicative of natural disasters. We validated RetNet across multiple disaster scenarios - landslides, floods, and wildfires - using images taken from various perspectives, including side-view, aerial and shipborne views. Our findings reveal an accuracy of 91.60% in landslide detection from side-view image captions and 87.50% for detections from shipborne views. These results underscore RetNet's effectiveness in enhancing the identification of disaster-affefted regions.
出版者	Institute of Electrical and Electronics Engineers Inc.
収録物名	IEEE Access
巻(号)	12
ページ	143754-143766
刊行年月	2024-10-11
言語	英語
ハンドルURL	https://hdl.handle.net/11094/99687
eISSN	21693536
関連情報 (isIdenticalTo)	DOI (出版社版) https://doi.org/10.1109/ACCESS.2024.3443130
アクセス権	open access
権利情報	This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
出版タイプ	Version of Record
カテゴリ	学術雑誌論文 Journal Article

資源タイプ	学術雑誌論文
ローカル資源タイプ	学術雑誌論文
DCTERMS.bibliographicCitation	IEEE Access. 2024, 12, p. 143754-143766
DC.title	Identifying Disaster Regions in Images Through Attention Shifting With a Retarget Network
DC.creator	Thanyawet, Narongthat
	Ratsamee, Photchara
	Uranishi, Yuki
	Kobayashi, Masato
	Takemura, Haruo
DC.publisher	Institute of Electrical and Electronics Engineers Inc.
DC.language' scheme='DCTERMS.RFC1766	英語
DCTERMS.issued' scheme='DCTERMS.W3CDTF	2024-10-11
DC.identifier	https://doi.org/10.1109/ACCESS.2024.3443130
DC.subject	anchor box
	disaster detection
	image captioning
	Retarget network
	vision transformer
DC.description	Thanyawet N., Ratsamee P., Uranishi Y., et al. Identifying Disaster Regions in Images Through Attention Shifting With a Retarget Network. IEEE Access 12, 143754 (2024); https://doi.org/10.1109/ACCESS.2024.3443130.
DCTERMS.abstract	Disasters disrupt lives and necessitate quick location of affected areas for rescue efforts. The application of computer vision has enhanced disaster detection, such as landslides and floods; however traditional computer vision methods often overlook smaller, critical details in favor of prominent objects. This research introduces the Retarget Network (RetNet), a novel framework aimed at improving image captioning techniques to identify and prioritize these less evident, yet crucial, objects in disaster scenarios, enhancing scene recognition and aiding in more effective disaster response. By masking images, we direct the model's focus towards additional significant areas within the image. RetNet employs anchor boxes to refine the targeting of specific areas and, optimize their center positions, heights, and widths. Additionally, RetNet determines which anchors to mask prior to captioning, facilitating the identification of challenging objects such as boulders, soil, and water, which are indicative of natural disasters. We validated RetNet across multiple disaster scenarios - landslides, floods, and wildfires - using images taken from various perspectives, including side-view, aerial and shipborne views. Our findings reveal an accuracy of 91.60% in landslide detection from side-view image captions and 87.50% for detections from shipborne views. These results underscore RetNet's effectiveness in enhancing the identification of disaster-affefted regions.
DC.rights	This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
DC.format	99687
citation_title	Identifying Disaster Regions in Images Through Attention Shifting With a Retarget Network
citation_author	Thanyawet, Narongthat
	Ratsamee, Photchara
	Uranishi, Yuki
	Kobayashi, Masato
	Takemura, Haruo
citation_publisher	Institute of Electrical and Electronics Engineers Inc.
citation_language	英語
citation_date	2024-10-11
citation_journal_title	IEEE Access
citation_volume	12
citation_firstpage	143754
citation_lastpage	143766
citation_doi	https://doi.org/10.1109/ACCESS.2024.3443130
citation_public_url	https://hdl.handle.net/11094/99687

Identifying Disaster Regions in Images Through Attention Shifting With a Retarget Network

閲覧可能ファイル

論文情報

論文詳細を表示