{"updated":"2025-01-19T17:33:24.973027+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00212269","sets":["1164:3616:10522:10655"]},"path":["10655"],"owner":"44499","recid":"212269","title":["ニューラルネットワークを用いた単一画像超解像のウェーブレット変換による高速化"],"pubdate":{"attribute_name":"公開日","attribute_value":"2021-08-18"},"_buckets":{"deposit":"6c5dc59c-fde8-473f-bd06-c8e82277b343"},"_deposit":{"id":"212269","pid":{"type":"depid","value":"212269","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"ニューラルネットワークを用いた単一画像超解像のウェーブレット変換による高速化","author_link":["541193","541192","541194","541195"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"ニューラルネットワークを用いた単一画像超解像のウェーブレット変換による高速化"}]},"item_type_id":"4","publish_date":"2021-08-18","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"防衛大学校理工学研究科"},{"subitem_text_value":"防衛大学校情報工学科"},{"subitem_text_value":"防衛大学校情報工学科"},{"subitem_text_value":"防衛大学校情報工学科"}]},"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/212269/files/IPSJ-AVM21114007.pdf","label":"IPSJ-AVM21114007.pdf"},"date":[{"dateType":"Available","dateValue":"2023-08-18"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-AVM21114007.pdf","filesize":[{"value":"629.7 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":"27"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"ce5c1742-49db-4ce1-8f03-4befb79adba3","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":"グェン, マントゥアン"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"岩井, 啓輔"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"松原, 隆"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"黒川, 恭一"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10438399","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":"2188-8582","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"本稿では，ニューラルネットワークを用いた単一画像超解像 (SISR，Single Image Super-Resolution) にウェーブレット変換 (WT，Wavelet Transform) を加えて高速化する手法を提案する．SISR は，一枚の低解像度画像から高解像度画像を作成する．機械学習の技術発展と共に，畳み込みニューラルネットワークを用いた EDSR (Enhanced Deep network SR) 等，高精度の SISR が提案されている．しかし，これらのネットワークは非常に階層が深く処理に時間がかかる．WT は，画像を 1 つの低周波成分と 3 つの高周波成分に分解し，WIT (Wavelet Inverse Transform) はその 4 成分を結合して元の画像に復元する．4 成分はサイズが元の画像サイズの 1/4 であり，スパース性を持つ．WT により生成された成分をニューラルネットワークに学習させる事で，画像生成時間や学習時間を大幅に下げる事が出来る．本稿では，WT を行う層 (WT Layer) と WIT を行う層 (WIT Layer) を既存の EDSR と組み合わせて，より高速な EDWSR (Enhanced Deep Wavelet network SR) を提案する．その結果，EDWSR は EDSR とほぼ同等の精度を保ちながら，約 3.6 倍の高速化を実現した．学習スピードも約 2 倍に高速化した．","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"6","bibliographic_titles":[{"bibliographic_title":"研究報告オーディオビジュアル複合情報処理（AVM）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2021-08-18","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"7","bibliographicVolumeNumber":"2021-AVM-114"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"created":"2025-01-19T01:13:14.930631+00:00","id":212269,"links":{}}