{"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00216972","sets":["1164:4619:10826:10881"]},"path":["10881"],"owner":"44499","recid":"216972","title":["UMAとMCDの組み合わせによる一人称視点動画におけるハンドセグメンテーション"],"pubdate":{"attribute_name":"公開日","attribute_value":"2022-03-03"},"_buckets":{"deposit":"cb9b1872-0c61-4177-9544-e4a76bf39a35"},"_deposit":{"id":"216972","pid":{"type":"depid","value":"216972","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"UMAとMCDの組み合わせによる一人称視点動画におけるハンドセグメンテーション","author_link":["561201","561202","561203"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"UMAとMCDの組み合わせによる一人称視点動画におけるハンドセグメンテーション"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"セッション6-A","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2022-03-03","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"大阪府立大学工学域"},{"subitem_text_value":"大阪府立大学大学院工学研究科"},{"subitem_text_value":"大阪府立大学大学院工学研究科"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"item_publisher":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"情報処理学会","subitem_publisher_language":"ja"}]},"publish_status":"0","weko_shared_id":-1,"item_file_price":{"attribute_name":"Billing file","attribute_type":"file","attribute_value_mlt":[{"url":{"url":"https://ipsj.ixsq.nii.ac.jp/record/216972/files/IPSJ-CVIM22229041.pdf","label":"IPSJ-CVIM22229041.pdf"},"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-CVIM22229041.pdf","filesize":[{"value":"3.2 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"0","billingrole":"20"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_login","version_id":"dad3b5b5-1cf8-4477-9735-7f0477edee09","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2022 by the Institute of Electronics, Information and Communication Engineers This SIG report is only available to those in membership of the SIG."}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"鈴木, 健一"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"井上, 勝文"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"吉岡, 理文"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AA11131797","subitem_source_identifier_type":"NCID"}]},"item_4_textarea_12":{"attribute_name":"Notice","attribute_value_mlt":[{"subitem_textarea_value":"SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc."}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_18gh","resourcetype":"technical report"}]},"item_4_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"2188-8701","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":" 一人称視点動画におけるドメインシフトは，学習データと異なる環境下の撮影時に，画像全体の外観の変化から生じやすく，ハンドセグメンテーションモデルの性能低下を招く．この問題に対して，既存手法である UMA は教師なしドメイン適応によって手領域特徴を自己学習することで精度を確保している．しかし，自己学習時に，ピクセルに対して誤った疑似ラベルが付与された場合の補正機能がなく，精度改善の余地がある．そこで本研究では，誤った疑似ラベル付与を防ぐために，敵対的学習によってドメイン適応する MCD を UMA に導入し，ハンドセグメンテーション性能の向上を図る．具体的には，手の疑似ラベルマスクのもととなる確率マップに関して，分類器によるドメイン予測の不一致に着目する．この不一致を用いた教師なしドメイン適応を行い，ターゲットデータの各ピクセルが手領域に含まれるかどうかを精度良く分類することで，誤った疑似ラベルの原因となる確率マップの生成を抑制する．公共データセットを用いた実験で，提案手法は UMA より，平均で約 3.8% の精度向上を達成した．","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"6","bibliographic_titles":[{"bibliographic_title":"研究報告コンピュータビジョンとイメージメディア（CVIM）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2022-03-03","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"41","bibliographicVolumeNumber":"2022-CVIM-229"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":216972,"updated":"2025-01-19T15:40:14.305475+00:00","links":{},"created":"2025-01-19T01:17:28.950343+00:00"}