{"links":{},"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00241067","sets":["1164:3206:11566:11793"]},"path":["11793"],"owner":"44499","recid":"241067","title":["パーリンノイズを用いたリモートセンシング画像に対する自己教師あり対照学習"],"pubdate":{"attribute_name":"公開日","attribute_value":"2024-11-22"},"_buckets":{"deposit":"36a48825-4b26-49af-9740-1214b5a20b9f"},"_deposit":{"id":"241067","pid":{"type":"depid","value":"241067","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"パーリンノイズを用いたリモートセンシング画像に対する自己教師あり対照学習","author_link":["663034","663035","663032","663033"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"パーリンノイズを用いたリモートセンシング画像に対する自己教師あり対照学習"},{"subitem_title":"Self-Supervised Contrastive Learning for Remote Sensing Imagery with Perlin Noise Injection","subitem_title_language":"en"}]},"item_type_id":"4","publish_date":"2024-11-22","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"東京電機大学"},{"subitem_text_value":"東京電機大学"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Tokyo Denki University","subitem_text_language":"en"},{"subitem_text_value":"Tokyo Denki University","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/241067/files/IPSJ-CG24196008.pdf","label":"IPSJ-CG24196008.pdf"},"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-CG24196008.pdf","filesize":[{"value":"1.2 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"0","billingrole":"28"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_login","version_id":"689cc5be-307e-45ab-ae1d-db132a40ad69","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2024 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":"岡澤, 律来"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"前田, 英作"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Riku, Okazawa","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Eisaku, Maeda","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10100541","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":"2188-8949","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"人工衛星の普及に伴いリモートセンシング（RS）画像の利活用が急速に進みつつあるが，RS 画像の解析には通常の画像認識タスクと同様に深層学習が有効であるとともに，タスクに応じた大量のラベル付き学習画像データが必要となる．限られた学習データを利用してタスクの性能を高めるためにラベルなしデータで事前学習を行うことが多いが，RS 画像にはセンサの性質や撮影時の状況によって品質にばらつきが生じるため，事前学習による十分な効果が得られない．そこで本研究では，RS 画像における画像品質のばらつきに着目し，入力画像にパーリンノイズを重ねるデータ拡張を対照学習に導入することを提案する．土地被覆分類タスクに本手法を適用し，パーリンノイズ導入前の手法と比較して本手法の有効性を確認した．また，より自然なノイズを入力画像に追加することで学習済みモデルの特徴量が土地利用分類において有益な特徴を示すことが示唆された．本手法は，RS 画像だけでなく画像品質に大きなばらつきがある画像全般に有効であると考えられる．","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"With the increasing use of satellites, the application of remote sensing (RS) imagery is rapidly advancing. Deep learning is effective for RS image analysis, but it requires large amounts of labeled data, which is often limited. To enhance performance, pre-training on unlabeled data is common. However, RS images vary in quality due to sensor characteristics and environmental conditions, reducing the effectiveness of pre-training. This study proposes the use of Perlin noise-based data augmentation in contrastive learning to address these variations. Transfered to land cover classification, the method improved performance and highlighted its potential for images with significant quality variations.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"6","bibliographic_titles":[{"bibliographic_title":"研究報告コンピュータグラフィックスとビジュアル情報学（CG）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2024-11-22","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"8","bibliographicVolumeNumber":"2024-CG-196"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"updated":"2025-01-19T07:45:16.006688+00:00","created":"2025-01-19T01:45:37.224320+00:00","id":241067}