{"created":"2025-01-19T01:12:44.008862+00:00","updated":"2025-01-19T17:44:55.771424+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00211588","sets":["1164:5159:10515:10608"]},"path":["10608"],"owner":"44499","recid":"211588","title":["多重解像度深層分析を用いた楽音分離の実験的評価"],"pubdate":{"attribute_name":"公開日","attribute_value":"2021-06-11"},"_buckets":{"deposit":"78a1f9a7-5423-4280-9087-2d069303d0b2"},"_deposit":{"id":"211588","pid":{"type":"depid","value":"211588","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"多重解像度深層分析を用いた楽音分離の実験的評価","author_link":["537790","537791"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"多重解像度深層分析を用いた楽音分離の実験的評価"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"一般発表","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2021-06-11","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"東京大学"},{"subitem_text_value":"東京大学"}]},"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/211588/files/IPSJ-SLP21137013.pdf","label":"IPSJ-SLP21137013.pdf"},"date":[{"dateType":"Available","dateValue":"2023-06-11"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-SLP21137013.pdf","filesize":[{"value":"1.3 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":"9f8b37fe-b819-43a8-9b4a-88718037a904","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":"中村, 友彦"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"猿渡, 洋"}],"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":"時間領域音源分離は時間周波数領域を介さず直接観測音響信号を処理し分離音を出力する技術であり，end-to-end 型の深層ニューラルネットワーク（deep neural network: DNN）を用いた手法が注目を集めている．Wave-U-Net は時間領域音源分離のための DNN の 1 つであり，特徴量を畳み込み層と非線形関数で処理しつつ繰り返しダウンサンプリングした後，入力と同一の時間解像度まで同様に繰り返しアップサンプリングを行う構造を持つ．しかし，Wave-U-Net のダウンサンプリング層はデシメーションにより実装されているため，特徴量領域でエイリアシングが起こるだけでなく，音源分離に有用な情報を含みうる特徴量の一部を破棄してしまう．これらの問題を同時に解決するため，我々は以前多重解像度解析と U-Net 構造の類似性に着眼し，離散ウェーブレット変換（discrete wavelet transform: DWT）に基づくダウンサンプリング層（DWT層）を用いた DNN ベース音源分離手法（多重解像度深層分析）を提案した．さらに，DWT 層に用いるウェーブレット基底関数を DNN と同時に学習できるよう拡張した．本稿では，これらの多重解像度深層分析とその拡張に対して詳細な検討を行うため，従来法との様々なモデルサイズでの比較を行う．この実験により，複数のモデルサイズにおいて多重解像度深層分析が従来法よりも高い分離性能を達成することを確認した．また，主観評価実験により，多重解像度深層分析が聴感上においても従来法より有意に高い分離性能を持つことを示した．","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"11","bibliographic_titles":[{"bibliographic_title":"研究報告音声言語情報処理（SLP）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2021-06-11","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"13","bibliographicVolumeNumber":"2021-SLP-137"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":211588,"links":{}}