{"created":"2025-01-18T23:30:56.041604+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00072658","sets":["1164:5159:6316:6317"]},"path":["6317"],"owner":"10","recid":"72658","title":["ウエーブレットの最適化と雑音プロファイルを用いた雑音抑圧による頑健な音声認識"],"pubdate":{"attribute_name":"公開日","attribute_value":"2011-01-28"},"_buckets":{"deposit":"722130ca-4c69-4693-83f0-bbaa301b857a"},"_deposit":{"id":"72658","pid":{"type":"depid","value":"72658","revision_id":0},"owners":[10],"status":"published","created_by":10},"item_title":"ウエーブレットの最適化と雑音プロファイルを用いた雑音抑圧による頑健な音声認識","author_link":["0","0"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"ウエーブレットの最適化と雑音プロファイルを用いた雑音抑圧による頑健な音声認識"},{"subitem_title":"Robust Speech Recognition Using Optimized Wavelet Denoising with Noise Profiles","subitem_title_language":"en"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"音声認識","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2011-01-28","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"京都大学学術情報メディアセンター"},{"subitem_text_value":"京都大学学術情報メディアセンター"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Academic Center for Computing and Media Studies (ACCMS), Kyoto University.","subitem_text_language":"en"},{"subitem_text_value":"Academic Center for Computing and Media Studies (ACCMS), Kyoto University.","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"eng"}]},"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/72658/files/IPSJ-SLP11085012.pdf"},"date":[{"dateType":"Available","dateValue":"2013-01-28"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-SLP11085012.pdf","filesize":[{"value":"198.5 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":"caa7f9ab-ff56-45d7-a32f-bdd3acf00860","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2011 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"ゴメス・ランディ"},{"creatorName":"河原, 達也"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Randy, Gomez","creatorNameLang":"en"},{"creatorName":"Tatsuya, Kawahara","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_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"本研究では、音声認識のためのウエーブレットに基づく雑音抑圧を雑音プロファイルと組み合わせることで改善を図る。学習時には、音声と種々の雑音プロファイル毎にウエーブレット変換のパラメータを最適化し、ウイナーフィルタのゲイン係数の推定の高精度化を図る。認識時には、雑音プロファイルを特定し、入力のウエーブレット係数を当該のウイナーゲインでフィルタリングする。さらに、ウイナーゲインにスケーリング係数を導入し、雑音抑圧に伴う歪みによるミスマッチを補償する。評価実験において、従来のウエーブレットに基づく手法と比較を行った。また、様々な雑音条件において頑健性の評価も行った。","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"In this paper, we improved the wavelet-based denoising method for automatic speech recognition (ASR) by using noise profiles. During training, we optimize the wavelet parameters for speech and different noise profiles to achieve a better estimate of the Wiener gain for effective filtering. Denoising is implemented by identifying the noise profile and filtering the noisy wavelet coefficients using a Wiener gain. In addition to wavelet filtering, we also introduce scale factors to the Wiener gain during decoding, to compensate for the mismatch caused by distortion during the denoising process. In our experimental evaluations, we compare our method with existing wavelet-based approach. We also conducted an experiment to test for robustness to different noise conditions.","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":"2011-01-28","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"12","bibliographicVolumeNumber":"2011-SLP-85"}]},"relation_version_is_last":true,"weko_creator_id":"10"},"links":{},"id":72658,"updated":"2025-01-21T22:40:48.954671+00:00"}