{"id":209752,"updated":"2025-01-19T18:24:08.911154+00:00","links":{},"created":"2025-01-19T01:11:02.466284+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00209752","sets":["1164:5159:10515:10530"]},"path":["10530"],"owner":"44499","recid":"209752","title":["雑音の基底信号を用いた耐雑音性の高い時間領域音声分離"],"pubdate":{"attribute_name":"公開日","attribute_value":"2021-02-24"},"_buckets":{"deposit":"e893442e-03e1-48e6-a045-cacee5c5f3a2"},"_deposit":{"id":"209752","pid":{"type":"depid","value":"209752","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"雑音の基底信号を用いた耐雑音性の高い時間領域音声分離","author_link":["529557","529559","529552","529558","529556","529555","529554","529553"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"雑音の基底信号を用いた耐雑音性の高い時間領域音声分離"},{"subitem_title":"Noise-robust time-domain speech separation with basis signals for noise","subitem_title_language":"en"}]},"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":"2021-02-24","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"東京工業大学情報理工学院"},{"subitem_text_value":"東京都市大学メディア情報学部"},{"subitem_text_value":"東京工業大学情報理工学院"},{"subitem_text_value":"東京工業大学情報理工学院"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Department of Computer Science, Tokyo Institute of Technology","subitem_text_language":"en"},{"subitem_text_value":"Faculty of Environmental Studies, Tokyo City University","subitem_text_language":"en"},{"subitem_text_value":"Department of Computer Science, Tokyo Institute of Technology","subitem_text_language":"en"},{"subitem_text_value":"Department of Computer Science, Tokyo Institute of Technology","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/209752/files/IPSJ-SLP21136014.pdf","label":"IPSJ-SLP21136014.pdf"},"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-SLP21136014.pdf","filesize":[{"value":"1.7 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"0","billingrole":"22"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_login","version_id":"d32e5ae8-2961-4488-91ae-81d057dbe39b","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2021 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":"尾座本, 耕平"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"岩野, 公司"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"宇都, 有昭"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"篠田, 浩一"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Kohei, Ozamoto","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Koji, Iwano","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Kuniaki, Uto","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Koichi, Sshinoda","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":"近年，深層学習を用いた音声分離が盛んに研究されている．波形を直接入力する時間領域の手法であるTasNetは，音声を畳み込みによって特徴量に変換した上で分離を行い，分離した特徴量に対して畳み込みを行って波形を再構成する．畳み込みフィルタは基底信号と呼ばれ，話者間の分離精度を高めるように学習される．この手法は，音声に雑音が含まれている場合に分離性能が大きく低下する．そこで我々は，話者の基底信号に加え雑音の基底信号を追加することで分離性能を向上させる方法 TasNet with noise basis signals（TasNet-NB）を提案する．雑音のSN比を徐々に小さくするカリキュラム学習と雑音の再構成損失の計算を用いる．WHAM!データセットを用いて評価した結果，SI-SDRiが13.7から14.6に向上した.","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"Recently, speech separation using deep learning has been extensively studied. TasNet, a time-domain method that directly inputs waveforms, converts speech into features by convolution, performs separation, and reconstructs waveform by convolution on separated features. The convolutional filter is called basis signals and is trained to improve separation accuracy between speakers. The separation performance of this method is greatly degraded when the speech contains noise. Therefore, we propose TasNet with noise basis signals (TasNet-NB), a method to improve separation performance by adding noise basis signals to speaker’s basis signals. We evaluate the method on WHAM! dataset and show that it improves SI-SDRi from 13.7 to 14.6.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"5","bibliographic_titles":[{"bibliographic_title":"研究報告音声言語情報処理（SLP）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2021-02-24","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"14","bibliographicVolumeNumber":"2021-SLP-136"}]},"relation_version_is_last":true,"weko_creator_id":"44499"}}