ログイン
言語:

WEKO3

  • トップ
  • コミュニティ
  • ランキング
AND
To
lat lon distance
To

Field does not validate



インデックスリンク

インデックスツリー

メールアドレスを入力してください。

WEKO

One fine body…

WEKO

One fine body…

アイテム

{"_buckets": {"deposit": "e029d2b3-35ba-4977-b8be-d592ce9ca29f"}, "_deposit": {"id": "26376", "owners": [], "pid": {"revision_id": 0, "type": "depid", "value": "26376"}, "status": "published"}, "_oai": {"id": "oai:nagoya.repo.nii.ac.jp:00026376"}, "item_10_biblio_info_6": {"attribute_name": "\u66f8\u8a8c\u60c5\u5831", "attribute_value_mlt": [{"bibliographicIssueDates": {"bibliographicIssueDate": "2018-06", "bibliographicIssueDateType": "Issued"}, "bibliographicPageEnd": "99", "bibliographicPageStart": "90", "bibliographicVolumeNumber": "66", "bibliographic_titles": [{"bibliographic_title": "Computerized Medical Imaging and Graphics"}]}]}, "item_10_description_4": {"attribute_name": "\u6284\u9332", "attribute_value_mlt": [{"subitem_description": "Recent advances in 3D fully convolutional networks (FCN) have made it feasible to produce dense voxel-wise predictions of volumetric images. In this work, we show that a multi-class 3D FCN trained on manually labeled CT scans of several anatomical structures (ranging from the large organs to thin vessels) can achieve competitive segmentation results, while avoiding the need for handcrafting features or training class-specific models. To this end, we propose a two-stage, coarse-to-fine approach that will first use a 3D FCN to roughly define a candidate region, which will then be used as input to a second 3D FCN. This reduces the number of voxels the second FCN has to classify to \u223c10% and allows it to focus on more detailed segmentation of the organs and vessels. We utilize training and validation sets consisting of 331 clinical CT images and test our models on a completely unseen data collection acquired at a different hospital that includes 150 CT scans, targeting three anatomical organs (liver, spleen, and pancreas). In challenging organs such as the pancreas, our cascaded approach improves the mean Dice score from 68.5 to 82.2%, achieving the highest reported average score on this dataset. We compare with a 2D FCN method on a separate dataset of 240 CT scans with 18 classes and achieve a significantly higher performance in small organs and vessels. Furthermore, we explore fine-tuning our models to different datasets. Our experiments illustrate the promise and robustness of current 3D FCN based semantic segmentation of medical images, achieving state-of-the-art results.", "subitem_description_type": "Abstract"}]}, "item_10_description_5": {"attribute_name": "\u5185\u5bb9\u8a18\u8ff0", "attribute_value_mlt": [{"subitem_description": "\u30d5\u30a1\u30a4\u30eb\u516c\u958b\uff1a2019-06-01", "subitem_description_type": "Other"}]}, "item_10_publisher_32": {"attribute_name": "\u51fa\u7248\u8005", "attribute_value_mlt": [{"subitem_publisher": "Elsevier"}]}, "item_10_relation_11": {"attribute_name": "DOI", "attribute_value_mlt": [{"subitem_relation_type_id": {"subitem_relation_type_id_text": "https://doi.org/10.1016/j.compmedimag.2018.03.001", "subitem_relation_type_select": "DOI"}}]}, "item_10_rights_12": {"attribute_name": "\u6a29\u5229", "attribute_value_mlt": [{"subitem_rights": "\u00a9 2018. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/"}]}, "item_10_select_15": {"attribute_name": "\u8457\u8005\u7248\u30d5\u30e9\u30b0", "attribute_value_mlt": [{"subitem_select_item": "author"}]}, "item_10_source_id_61": {"attribute_name": "ISSN\uff08print\uff09", "attribute_value_mlt": [{"subitem_source_identifier": "0895-6111", "subitem_source_identifier_type": "ISSN"}]}, "item_creator": {"attribute_name": "\u8457\u8005", "attribute_type": "creator", "attribute_value_mlt": [{"creatorNames": [{"creatorName": "Roth, Holger R."}], "nameIdentifiers": [{"nameIdentifier": "85339", "nameIdentifierScheme": "WEKO"}]}, {"creatorNames": [{"creatorName": "Oda, Hirohisa"}], "nameIdentifiers": [{"nameIdentifier": "85340", "nameIdentifierScheme": "WEKO"}]}, {"creatorNames": [{"creatorName": "Zhou, Xiangrong"}], "nameIdentifiers": [{"nameIdentifier": "85341", "nameIdentifierScheme": "WEKO"}]}, {"creatorNames": [{"creatorName": "Shimizu, Natsuki"}], "nameIdentifiers": [{"nameIdentifier": "85342", "nameIdentifierScheme": "WEKO"}]}, {"creatorNames": [{"creatorName": "Yang, Ying"}], "nameIdentifiers": [{"nameIdentifier": "85343", "nameIdentifierScheme": "WEKO"}]}, {"creatorNames": [{"creatorName": "Hayashi, Yuichiro"}], "nameIdentifiers": [{"nameIdentifier": "85344", "nameIdentifierScheme": "WEKO"}]}, {"creatorNames": [{"creatorName": "Oda, Masahiro"}], "nameIdentifiers": [{"nameIdentifier": "85345", "nameIdentifierScheme": "WEKO"}]}, {"creatorNames": [{"creatorName": "Fujiwara, Michitaka"}], "nameIdentifiers": [{"nameIdentifier": "85346", "nameIdentifierScheme": "WEKO"}]}, {"creatorNames": [{"creatorName": "Misawa, Kazunari"}], "nameIdentifiers": [{"nameIdentifier": "85347", "nameIdentifierScheme": "WEKO"}]}, {"creatorNames": [{"creatorName": "Mori, Kensaku"}], "nameIdentifiers": [{"nameIdentifier": "85348", "nameIdentifierScheme": "WEKO"}]}]}, "item_files": {"attribute_name": "\u30d5\u30a1\u30a4\u30eb\u60c5\u5831", "attribute_type": "file", "attribute_value_mlt": [{"accessrole": "open_date", "date": [{"dateType": "Available", "dateValue": "2019-06-01"}], "displaytype": "detail", "download_preview_message": "", "file_order": 0, "filename": "cmig2018_rothhr_cascaded_fcn_final.pdf", "filesize": [{"value": "10.0 MB"}], "format": "application/pdf", "future_date_message": "", "is_thumbnail": false, "licensetype": "license_free", "mimetype": "application/pdf", "size": 10000000.0, "url": {"label": "cmig2018_rothhr_cascaded_fcn_final", "url": "https://nagoya.repo.nii.ac.jp/record/26376/files/cmig2018_rothhr_cascaded_fcn_final.pdf"}, "version_id": "0522049c-cca5-4d54-a67c-9cbce9f9960b"}]}, "item_keyword": {"attribute_name": "\u30ad\u30fc\u30ef\u30fc\u30c9", "attribute_value_mlt": [{"subitem_subject": "Fully convolutional networks", "subitem_subject_scheme": "Other"}, {"subitem_subject": "Deep learning", "subitem_subject_scheme": "Other"}, {"subitem_subject": "Medical imaging", "subitem_subject_scheme": "Other"}, {"subitem_subject": "Computed tomography", "subitem_subject_scheme": "Other"}, {"subitem_subject": "Multi-organ segmentation", "subitem_subject_scheme": "Other"}]}, "item_language": {"attribute_name": "\u8a00\u8a9e", "attribute_value_mlt": [{"subitem_language": "eng"}]}, "item_resource_type": {"attribute_name": "\u8cc7\u6e90\u30bf\u30a4\u30d7", "attribute_value_mlt": [{"resourcetype": "journal article", "resourceuri": "http://purl.org/coar/resource_type/c_6501"}]}, "item_title": "An application of cascaded 3D fully convolutional networks for medical image segmentation", "item_titles": {"attribute_name": "\u30bf\u30a4\u30c8\u30eb", "attribute_value_mlt": [{"subitem_title": "An application of cascaded 3D fully convolutional networks for medical image segmentation"}]}, "item_type_id": "10", "owner": "1", "path": ["312/313/314"], "permalink_uri": "http://hdl.handle.net/2237/00028579", "pubdate": {"attribute_name": "\u516c\u958b\u65e5", "attribute_value": "2018-09-18"}, "publish_date": "2018-09-18", "publish_status": "0", "recid": "26376", "relation": {}, "relation_version_is_last": true, "title": ["An application of cascaded 3D fully convolutional networks for medical image segmentation"], "weko_shared_id": null}
  1. A500 情報学部/情報学研究科・情報文化学部・情報科学研究科
  2. A500a 雑誌掲載論文
  3. 学術雑誌

An application of cascaded 3D fully convolutional networks for medical image segmentation

http://hdl.handle.net/2237/00028579
908f5089-907c-426b-8b3a-3c140d274e38
名前 / ファイル ライセンス アクション
cmig2018_rothhr_cascaded_fcn_final.pdf cmig2018_rothhr_cascaded_fcn_final (10.0 MB)
Item type 学術雑誌論文 / Journal Article(1)
公開日 2018-09-18
タイトル
タイトル An application of cascaded 3D fully convolutional networks for medical image segmentation
著者 Roth, Holger R.

× Roth, Holger R.

WEKO 85339

Roth, Holger R.

Search repository
Oda, Hirohisa

× Oda, Hirohisa

WEKO 85340

Oda, Hirohisa

Search repository
Zhou, Xiangrong

× Zhou, Xiangrong

WEKO 85341

Zhou, Xiangrong

Search repository
Shimizu, Natsuki

× Shimizu, Natsuki

WEKO 85342

Shimizu, Natsuki

Search repository
Yang, Ying

× Yang, Ying

WEKO 85343

Yang, Ying

Search repository
Hayashi, Yuichiro

× Hayashi, Yuichiro

WEKO 85344

Hayashi, Yuichiro

Search repository
Oda, Masahiro

× Oda, Masahiro

WEKO 85345

Oda, Masahiro

Search repository
Fujiwara, Michitaka

× Fujiwara, Michitaka

WEKO 85346

Fujiwara, Michitaka

Search repository
Misawa, Kazunari

× Misawa, Kazunari

WEKO 85347

Misawa, Kazunari

Search repository
Mori, Kensaku

× Mori, Kensaku

WEKO 85348

Mori, Kensaku

Search repository
権利
権利情報 © 2018. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/
キーワード
主題Scheme Other
主題 Fully convolutional networks
キーワード
主題Scheme Other
主題 Deep learning
キーワード
主題Scheme Other
主題 Medical imaging
キーワード
主題Scheme Other
主題 Computed tomography
キーワード
主題Scheme Other
主題 Multi-organ segmentation
抄録
内容記述 Recent advances in 3D fully convolutional networks (FCN) have made it feasible to produce dense voxel-wise predictions of volumetric images. In this work, we show that a multi-class 3D FCN trained on manually labeled CT scans of several anatomical structures (ranging from the large organs to thin vessels) can achieve competitive segmentation results, while avoiding the need for handcrafting features or training class-specific models. To this end, we propose a two-stage, coarse-to-fine approach that will first use a 3D FCN to roughly define a candidate region, which will then be used as input to a second 3D FCN. This reduces the number of voxels the second FCN has to classify to ∼10% and allows it to focus on more detailed segmentation of the organs and vessels. We utilize training and validation sets consisting of 331 clinical CT images and test our models on a completely unseen data collection acquired at a different hospital that includes 150 CT scans, targeting three anatomical organs (liver, spleen, and pancreas). In challenging organs such as the pancreas, our cascaded approach improves the mean Dice score from 68.5 to 82.2%, achieving the highest reported average score on this dataset. We compare with a 2D FCN method on a separate dataset of 240 CT scans with 18 classes and achieve a significantly higher performance in small organs and vessels. Furthermore, we explore fine-tuning our models to different datasets. Our experiments illustrate the promise and robustness of current 3D FCN based semantic segmentation of medical images, achieving state-of-the-art results.
内容記述タイプ Abstract
内容記述
内容記述 ファイル公開:2019-06-01
内容記述タイプ Other
出版者
出版者 Elsevier
言語
言語 eng
資源タイプ
資源タイプresource http://purl.org/coar/resource_type/c_6501
タイプ journal article
DOI
関連識別子
識別子タイプ DOI
関連識別子 https://doi.org/10.1016/j.compmedimag.2018.03.001
ISSN(print)
収録物識別子タイプ ISSN
収録物識別子 0895-6111
書誌情報 Computerized Medical Imaging and Graphics

巻 66, p. 90-99, 発行日 2018-06
著者版フラグ
値 author
戻る
0
views
See details
Views

Versions

Ver.1 2021-03-01 11:24:41.622417
Show All versions

Share

Mendeley CiteULike Twitter Facebook Print Addthis

Cite as

Export

OAI-PMH
  • OAI-PMH JPCOAR
  • OAI-PMH DublinCore
  • OAI-PMH DDI
Other Formats
  • JSON
  • BIBTEX

Confirm


Powered by CERN Data Centre & Invenio


Powered by CERN Data Centre & Invenio