{"id":189982,"updated":"2025-01-20T01:25:15.564448+00:00","links":{},"created":"2025-01-19T00:55:58.702045+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00189982","sets":["1164:5159:9402:9506"]},"path":["9506"],"owner":"11","recid":"189982","title":["von Mises分布DNNに基づく振幅スペクトログラムからの位相復元"],"pubdate":{"attribute_name":"公開日","attribute_value":"2018-06-09"},"_buckets":{"deposit":"00af8969-ba54-41ca-9cf8-e021af702520"},"_deposit":{"id":"189982","pid":{"type":"depid","value":"189982","revision_id":0},"owners":[11],"status":"published","created_by":11},"item_title":"von Mises分布DNNに基づく振幅スペクトログラムからの位相復元","author_link":["433853","433857","433858","433855","433859","433861","433860","433856","433862","433854"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"von Mises分布DNNに基づく振幅スペクトログラムからの位相復元"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"ポスターセッション","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2018-06-09","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"東京大学大学院情報理工学系システム情報学専攻"},{"subitem_text_value":"東京大学大学院情報理工学系システム情報学専攻"},{"subitem_text_value":"東京大学大学院情報理工学系システム情報学専攻"},{"subitem_text_value":"香川高等専門学校電気情報工学科"},{"subitem_text_value":"東京大学大学院情報理工学系システム情報学専攻"}]},"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/189982/files/IPSJ-SLP18122054.pdf","label":"IPSJ-SLP18122054.pdf"},"date":[{"dateType":"Available","dateValue":"2020-06-09"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-SLP18122054.pdf","filesize":[{"value":"792.1 kB"}],"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":"fd539393-ef83-401a-9838-9cbd8b5d0ad6","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2018 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"高道, 慎之介"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"齋藤, 佑樹"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"高宗, 典玄"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"北村, 大地"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"猿渡, 洋"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Shinnosuke, Takamichi","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Yuki, Saito","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Norihiro, Takamune","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Daichi, Kitamura","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Hiroshi, Saruwatari","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"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":"本稿では，deep neural network (DNN) に基づく振幅スペクトログラムからの位相復元について述べる．音声音響信号処理では振幅スペクトログラムに対する処理がしばしば行われ，その位相スペクトログラムは得られない場合が多い．これに対し Griffin-Lim 法は，無矛盾性に基づき振幅スペクトログラムから位相を復元するが，生成音声に対して不自然なアーティファクトをもたらす．この問題に対処するために，本論文では von Mises 分布 DNN を導入する．この DNN は，位相のような周期変数の確率密度関数である von Mises 分布を条件付き分布として有する深層生成モデルであり，そのモデルパラメータは最尤基準で学習される．我々は，これを振幅スペクトログラムからの位相復元に適用し，更に，推定された位相の群遅延を自然な群遅延に近づけるための DNN 学習基準を導入する．実験結果より，(1) DNN は，位相そのものより群遅延を高精度に推定できること，また，(2) 提案法は，従来の Griffin-Lim 法を超える音質を達成できることを示す．","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":"2018-06-09","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"54","bibliographicVolumeNumber":"2018-SLP-122"}]},"relation_version_is_last":true,"weko_creator_id":"11"}}