{"created":"2021-03-01T06:16:38.790310+00:00","id":9854,"links":{},"metadata":{"_buckets":{"deposit":"6e5203a8-ab52-4aa7-8933-a69e819bb266"},"_deposit":{"id":"9854","owners":[],"pid":{"revision_id":0,"type":"depid","value":"9854"},"status":"published"},"_oai":{"id":"oai:nagoya.repo.nii.ac.jp:00009854","sets":["312:651:734"]},"author_link":["29683","29684"],"item_12_biblio_info_6":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicIssueDates":{"bibliographicIssueDate":"2009-03-25","bibliographicIssueDateType":"Issued"}}]},"item_12_date_granted_64":{"attribute_name":"学位授与年月日","attribute_value_mlt":[{"subitem_dategranted":"2009-03-25"}]},"item_12_degree_grantor_62":{"attribute_name":"学位授与機関","attribute_value_mlt":[{"subitem_degreegrantor":[{"subitem_degreegrantor_language":"ja","subitem_degreegrantor_name":"名古屋大学"},{"subitem_degreegrantor_language":"en","subitem_degreegrantor_name":"Nagoya University"}],"subitem_degreegrantor_identifier":[{"subitem_degreegrantor_identifier_name":"13901","subitem_degreegrantor_identifier_scheme":"kakenhi"}]}]},"item_12_degree_name_61":{"attribute_name":"学位名","attribute_value_mlt":[{"subitem_degreename":"博士(情報科学)","subitem_degreename_language":"ja"}]},"item_12_description_4":{"attribute_name":"抄録","attribute_value_mlt":[{"subitem_description":"In this thesis, a training method for the low-resolution character recognition task is proposed. It is named “Generative learning method,” since the training images are generated artificially from an original image. The generative learning method is applied to camera-based character recognition and traffic sign recognition. Recognition technologies using digital cameras have gained considerable interest in recent years. Provided that these technologies come into practical use, there could be useful applications for camera-equipped devices. However, even with the improvements of digital cameras, the quality of captured images is still insufficient for the recognition in many practical cases. This work focuses on the training of image degradation characteristics. Recognition methods presented in this thesis are based on the generative learning method in which training images are artificially generated. Conventional approaches used camera-captured images as training data, which required exhaustive collection of the sample images by actual capturing. The proposed generative learning method, instead, allows to obtain these training images based on a small set of actual images. Since the training images need to be generated on the basis of actual degradation characteristics, the estimation step of the degradation characteristics is introduced. This framework is applied to three applications . character recognition, text recognition, and traffic sign recognition. In the camera-based character recognition application, optical blur and the vibration of hand-held cameras seriously affect the recognition accuracy. The proposed method copes with the optical blur and the motion blur by generating degraded training images. They are generated using a PSF (Point Spread Function) that preserves the actual blur characteristics. In the text recognition application, segmentation of characters is an unavoidable problem. It is difficult especially in the case where the image resolution is low. The proposed method copes with this problem by introducing a segmentation model to the generative learning method. In addition to characters, patterns of spaces between two adjacent characters are generated and used for accurate segmentation of the characters. In the traffic sign recognition application, various factors influence the captured traffic sign images. In the proposed framework, degradation parameters are defined for the simulation of these degradation factors. The distribution of the degradation parameters are estimated from actual images because it can produce appropriate parameters for generating the training images. Results obtained here have proven that the proposed generative learning method is effective for the applications suffering from various image degradations.","subitem_description_language":"en","subitem_description_type":"Abstract"}]},"item_12_description_5":{"attribute_name":"内容記述","attribute_value_mlt":[{"subitem_description":"名古屋大学博士学位論文 学位の種類:博士(情報科学) (課程) 学位授与年月日:平成21年3月25日","subitem_description_language":"ja","subitem_description_type":"Other"}]},"item_12_dissertation_number_65":{"attribute_name":"学位授与番号","attribute_value_mlt":[{"subitem_dissertationnumber":"甲第8396号"}]},"item_12_identifier_60":{"attribute_name":"URI","attribute_value_mlt":[{"subitem_identifier_type":"HDL","subitem_identifier_uri":"http://hdl.handle.net/2237/11663"}]},"item_12_select_15":{"attribute_name":"著者版フラグ","attribute_value_mlt":[{"subitem_select_item":"publisher"}]},"item_12_text_14":{"attribute_name":"フォーマット","attribute_value_mlt":[{"subitem_text_value":"application/pdf"}]},"item_12_text_63":{"attribute_name":"学位授与年度","attribute_value_mlt":[{"subitem_text_value":"2008"}]},"item_access_right":{"attribute_name":"アクセス権","attribute_value_mlt":[{"subitem_access_right":"open access","subitem_access_right_uri":"http://purl.org/coar/access_right/c_abf2"}]},"item_creator":{"attribute_name":"著者","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Ishida, Hiroyuki","creatorNameLang":"en"}],"nameIdentifiers":[{"nameIdentifier":"29683","nameIdentifierScheme":"WEKO"}]},{"creatorNames":[{"creatorName":"石田, 皓之","creatorNameLang":"ja"}],"nameIdentifiers":[{"nameIdentifier":"29684","nameIdentifierScheme":"WEKO"}]}]},"item_files":{"attribute_name":"ファイル情報","attribute_type":"file","attribute_value_mlt":[{"accessrole":"open_date","date":[{"dateType":"Available","dateValue":"2018-02-20"}],"displaytype":"detail","filename":"thesis-ishida.pdf","filesize":[{"value":"1.7 MB"}],"format":"application/pdf","licensetype":"license_note","mimetype":"application/pdf","url":{"label":"thesis-ishida.pdf","objectType":"fulltext","url":"https://nagoya.repo.nii.ac.jp/record/9854/files/thesis-ishida.pdf"},"version_id":"d27ca44d-e30e-44cb-9a4d-55b7717ff1d3"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"eng"}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourcetype":"doctoral thesis","resourceuri":"http://purl.org/coar/resource_type/c_db06"}]},"item_title":"A generative learning method for low-resolution character recognition","item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"A generative learning method for low-resolution character recognition","subitem_title_language":"en"}]},"item_type_id":"12","owner":"1","path":["734"],"pubdate":{"attribute_name":"PubDate","attribute_value":"2009-05-01"},"publish_date":"2009-05-01","publish_status":"0","recid":"9854","relation_version_is_last":true,"title":["A generative learning method for low-resolution character recognition"],"weko_creator_id":"1","weko_shared_id":-1},"updated":"2023-01-16T04:41:05.150846+00:00"}