{"created":"2025-01-18T22:47:30.782405+00:00","updated":"2025-01-23T01:14:38.491435+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00013424","sets":["581:729:738"]},"path":["738"],"owner":"1","recid":"13424","title":["K - M木探索の範囲狭化による近傍文字識別の高速化"],"pubdate":{"attribute_name":"公開日","attribute_value":"1997-04-15"},"_buckets":{"deposit":"f6c84f95-044e-4501-8ffb-acf5150748cb"},"_deposit":{"id":"13424","pid":{"type":"depid","value":"13424","revision_id":0},"owners":[1],"status":"published","created_by":1},"item_title":"K - M木探索の範囲狭化による近傍文字識別の高速化","author_link":["0","0"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"K - M木探索の範囲狭化による近傍文字識別の高速化"},{"subitem_title":"Speed -up of a Nearest- neighbor Character Classifier by Narrowing the Search Area over a K - M Tree","subitem_title_language":"en"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"論文","subitem_subject_scheme":"Other"}]},"item_type_id":"2","publish_date":"1997-04-15","item_2_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"グローリー工業株式会社中央研究所"},{"subitem_text_value":"中央大学理工学部情報工学科"}]},"item_2_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Central Research Laboratory, GLORY Ltd","subitem_text_language":"en"},{"subitem_text_value":"Department of Information and System Engineering, Faculty of Science and Engineering, Chuo University","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"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/13424/files/IPSJ-JNL3804010.pdf"},"date":[{"dateType":"Available","dateValue":"1999-04-15"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-JNL3804010.pdf","filesize":[{"value":"1.1 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"660","billingrole":"5"},{"tax":["include_tax"],"price":"330","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"8"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"cf4ad570-d6d9-4a7b-9ae1-0f1f2b82ede3","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 1997 by the Information Processing Society of Japan"}]},"item_2_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"亀山, 博史"},{"creatorName":"鈴木, 寿"}],"nameIdentifiers":[{}]}]},"item_2_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Hirofumi, Kameyama","creatorNameLang":"en"},{"creatorName":"Hisashi, Suzuki","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_2_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN00116647","subitem_source_identifier_type":"NCID"}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_6501","resourcetype":"journal article"}]},"item_2_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"1882-7764","subitem_source_identifier_type":"ISSN"}]},"item_2_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"文字認識の実用化において，識別系の設計を効率良く行うことは重要である．最近傍識別は，識別能力が高く辞書の作成が容易であるという特長を持っているため，実用的に有効であると考えられる．しかし1つの難点として，参照パターンとの照合をいかに高速に行うかという問題がある．本論文では，K?M木とよばれる2分木を辞書のデータ構造として用い，木内の探索における距離計算の回数を削減する手法について述べる．K?M木へのデータの格納は高速に行えるため，多様な読み取り対象に対応して高性能な識別系を迅速に設計できることが期待される．本手法では，従来の三角不等式に基づいて探索範囲を狭化する条件にパラメータαを導入することにより，探索範囲をより狭化できるようにした．数字，英大文字およびカタカナからなる4つの手書き文字のサンプルを対象とした実験を行った結果，適切にパラメータαを設定することによって，高い正読率を保ちながら大幅な高速化が達成され，実用的に有効であることが示された．また，このようなパラメータの設定方法についても検討した．","subitem_description_type":"Other"}]},"item_2_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"The nearest-neighbor classifier is effective for the practical use of character recognition,since its ability of discrimination is high and a discrimination dictionary can be made easily.However,it is a problem to speed up the search in a dictionary for the nearest pattern.This paper presents a mathod of eliminating the number of distance computations for the nearestneighbor character classifier by using a K-M tree which is a binary tree proposed by Kalantari and McDonald.Since construction of a K-M tree is quickly executed,it can be expected that a high-performance classifier will be efficiently designed according to various character sets.In searching the nearest pattern over the K-M tree,the search area is narrowed by using a rule based on a triangle inequality.For further narrowing of the search area over the K-M tree,a parameter α is introduced into the rule.Experiments were made by using four kinds of hand-written character sets(two numeral sets,a capital alphabet set,and a katakana set).The results suggest a practical efficiency of the method:The classifier with a parameter α chosen appropriately keeps the correct recognition rate high,and a good number of distance computations are eliminated.A method of setting the parameter appropriately has also been presented.","subitem_description_type":"Other"}]},"item_2_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"770","bibliographic_titles":[{"bibliographic_title":"情報処理学会論文誌"}],"bibliographicPageStart":"762","bibliographicIssueDates":{"bibliographicIssueDate":"1997-04-15","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"4","bibliographicVolumeNumber":"38"}]},"relation_version_is_last":true,"item_2_alternative_title_2":{"attribute_name":"その他タイトル","attribute_value_mlt":[{"subitem_alternative_title":"画像・図形認識"}]},"weko_creator_id":"1"},"id":13424,"links":{}}