{"created":"2025-01-19T01:12:21.667774+00:00","updated":"2025-01-19T17:53:12.122498+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00211173","sets":["1164:4619:10416:10591"]},"path":["10591"],"owner":"44499","recid":"211173","title":["部分欠落および観測視点変化を含む歩容認証の頑健性評価"],"pubdate":{"attribute_name":"公開日","attribute_value":"2021-05-13"},"_buckets":{"deposit":"87c0cee1-6635-4850-9a5c-570a0b677324"},"_deposit":{"id":"211173","pid":{"type":"depid","value":"211173","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"部分欠落および観測視点変化を含む歩容認証の頑健性評価","author_link":["536084","536085","536083"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"部分欠落および観測視点変化を含む歩容認証の頑健性評価"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"卒論スポットライトセッション","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2021-05-13","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"大阪大学産業科学研究所"},{"subitem_text_value":"大阪大学産業科学研究所"},{"subitem_text_value":"大阪大学産業科学研究所"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"The Institute of Scientific and Industrial Research, Osaka Univ.","subitem_text_language":"en"},{"subitem_text_value":"The Institute of Scientific and Industrial Research, Osaka Univ.","subitem_text_language":"en"},{"subitem_text_value":"The Institute of Scientific and Industrial Research, Osaka Univ.","subitem_text_language":"en"}]},"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/211173/files/IPSJ-CVIM21226026.pdf","label":"IPSJ-CVIM21226026.pdf"},"date":[{"dateType":"Available","dateValue":"2023-05-13"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-CVIM21226026.pdf","filesize":[{"value":"1.8 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"660","billingrole":"5"},{"tax":["include_tax"],"price":"330","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"20"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"390ab045-bd19-48dc-80e1-45b68fd72c13","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2021 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"辻, 翔吾"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"槇原, 靖"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"八木, 康史"}],"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":"本研究では，映像を入力として，周囲にいる人物が事前に登録された歩行映像と同一人物かどうかを判定するための歩容認証に取り組んだ．このような歩容認証では，対象人物がカメラに向かって歩いてくる方向（観測視点）が登録・照合間で変化するため，観測される歩行映像やそこから抽出される歩容特徴が大きく変動する．加えて，入力において，他の物体に遮られていたり，視界からフレームアウトしていたりすることで，対象人物の全身像を得られないことがあるという問題もある．そこで，本研究では，観測視点の変化に頑健な歩容認証手法の一つである空間位置合わせネットワークと認識ネットワークを組み合わせた深層学習モデルを用いて，全身像が映らないことによる部分欠落や観測視点変化による影響を評価した．実験では，世界最大規模の歩行映像データベースである OUMVLP を用いて訓練・評価を行った．部分欠落の割合は，対象人物の上端から最大 60% までを対象とし，欠落を含めた入力と，欠落部分を抜き，観測可能な部分が上端に接するよう引き延ばした入力の二種類に対して訓練を行った．観測視点の変化は，対象人物はカメラに対し向かって歩いてくるという前提に立ち，正面（0°）を基本として，15°，30°を対象とした．評価は，これらのモデル・観測視点の変化に対し，他人誤受入率と本人誤拒否率の等価誤り率，および他人誤受入率が 1% の時の本人拒否率の 2 つの指標を用いて行った．結果として，本人誤拒否率の上限を 5% としたとき，登録時の観測方向が15°であれば，欠落割合が 30% まで許容できることが明らかになった．","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"8","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":"2021-05-13","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"26","bibliographicVolumeNumber":"2021-CVIM-226"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":211173,"links":{}}