{"links":{},"id":2002454,"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:02002454","sets":["1164:5159:1739855613862:1748588592959"]},"path":["1748588592959"],"owner":"80578","recid":"2002454","title":["マルチチャネルなりすまし音声検出における自己教師あり学習フロントエンドの有効性検証"],"pubdate":{"attribute_name":"PubDate","attribute_value":"2025-06-06"},"_buckets":{"deposit":"6f7fda24-92bc-49b2-923d-b5e5d92fd9ba"},"_deposit":{"id":"2002454","pid":{"type":"depid","value":"2002454","revision_id":0},"owners":[80578],"status":"published","created_by":80578},"item_title":"マルチチャネルなりすまし音声検出における自己教師あり学習フロントエンドの有効性検証","author_link":[],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"マルチチャネルなりすまし音声検出における自己教師あり学習フロントエンドの有効性検証","subitem_title_language":"ja"}]},"item_type_id":"4","publish_date":"2025-06-06","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"東京都立大学大学院システムデザイン研究科"},{"subitem_text_value":"日本電信電話株式会社"},{"subitem_text_value":"東京都立大学大学院システムデザイン研究科"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Tokyo Metropolitan University","subitem_text_language":"en"},{"subitem_text_value":"NTT","subitem_text_language":"en"},{"subitem_text_value":"Tokyo Metropolitan University","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/2002454/files/IPSJ-SLP25156064.pdf","label":"IPSJ-SLP25156064.pdf"},"date":[{"dateType":"Available","dateValue":"2027-06-06"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-SLP25156064.pdf","filesize":[{"value":"1.1 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":"22"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"2ff884b2-142d-4958-acac-e5ffc40d1079","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2025 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"山口,拓生"}]},{"creatorNames":[{"creatorName":"俵,直弘"}]},{"creatorNames":[{"creatorName":"塩田,さやか"}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10442647","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":"2188-8663","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"音声によって操作を行うデバイスの普及に伴い，音声認証技術の重要性が高まっている．しかし音声認証はなりすまし音声攻撃に対して脆弱であることが知られており，中でも録音再生によるなりすまし音声攻撃は検出が難しく，深刻な課題となっている．音声操作デバイスには通常，複数のマイクロフォンが搭載されており，マルチチャネル情報を利用することで空間的な特徴を活用した検出が可能となるが，これまでのなりすまし音声検出に関する研究は主にシングルチャネル音声を対象としており，マルチチャネル情報の有効性については十分に検討されてこなかった．そこで本研究では，マルチチャネル情報を活用することで，なりすまし音声の検出精度を向上させる方法について検討する．具体的には，フロントエンド処理として信号処理に基づく手法と自己教師あり学習（self supervised learning; SSL）に基づく手法の2種類を使用し，それぞれを異なる2種類のバックエンド識別モデルと組み合わせて評価した．公開されているマルチチャネルなりすまし音声コーパスを用いた実験の結果，SSLに基づくの特徴量を用いた手法は，従来手法と比較して等価エラー率を約14％低減し，検出精度が大幅に向上することが確認された．","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"6","bibliographic_titles":[{"bibliographic_title":"研究報告音声言語情報処理（SLP）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2025-06-06","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"64","bibliographicVolumeNumber":"2025-SLP-156"}]},"relation_version_is_last":true,"weko_creator_id":"80578"},"created":"2025-05-30T07:39:00.537987+00:00","updated":"2025-05-30T07:39:04.831461+00:00"}