Studies on efficient methods for large-size bibliographic data conversion.

Research Project

Project/Area Number	62880006
Research Category	Grant-in-Aid for Developmental Scientific Research
Allocation Type	Single-year Grants
Research Field	Informatics
Research Institution	THE UNIVERSITY OF TOKYO
Principal Investigator	KURODA Haruo (1988-1989) The University of Tokyo, Faculty of Science Professor, 理学部, 教授 (00011479) 山崎弘郎 (1987) 東京大学, 工学部, 教授
Co-Investigator(Kenkyū-buntansha)	YASUNAGA Hisashi Research Institute of Japanese Literature Professor, 教授 (20017411) OHYAMA Keizo National Center for Science Information System Associate Professor, 助教授 (90177022) MIYAZAWA Akira National Center for Science Information System Associate Professor, 助教授 (80099928) INOSE Hiroshi National Center for Science Information System Director General, 所長 (70010618) YAMASAKI Hiro The University of Tokyo, Faculty of Science Professor, 工学部, 教授 (30092365)
Project Period (FY)	1987 – 1989
Project Status	Completed (Fiscal Year 1989)
Budget Amount *help	¥40,200,000 (Direct Cost: ¥40,200,000) Fiscal Year 1989: ¥2,600,000 (Direct Cost: ¥2,600,000) Fiscal Year 1988: ¥20,600,000 (Direct Cost: ¥20,600,000) Fiscal Year 1987: ¥17,000,000 (Direct Cost: ¥17,000,000)
Keywords	SPEECH RECOGNITION / OPTICAL CHARACTER RECOGNITION / AUTOMATED DATA EXTRACTION / KNOWLEDGE BASE / UNION CATALOG DATABASE / RETROSPECTIVE CONVERSION / OPTICAL DISKS / CD-ROM / 文字認識 / フィ-ルド自動抽出・認識 / 音声入力 / 画像入力 / CDーROM / 音声認識 / 文字切り出し処理 / 自動認識 / 理解システム / 知識処理 / 進化的プログラミング環境 / オンライン入力 / 学術情報センター / ダウンロード / アップロード
Research Abstract	The cost of retrospective conversion of cord catalogs remains a substantial expense and a hindrance to availability of whole collection of nationwide academic information resources. The aim of this research project is to find out the ways to automate or improve the two phases of the ordinary retrospective conversion procedures. One is to prepare the computer readable data primarily from the catalog cards, and the other is to identify and integrate into the database in the context of the NACSIS online union catalog database. Evaluation of the availability of a conventional optical disk storage system is also considered. The research was carried out under the collaboration of librarians and information scientists belonging 8 national universities and 2 research institutes. The results obtained are as follows; 1) A character recognition system for catalog cards, in which a speaker-dependent voice recognition or optical scanning is used, was developed. The system correctly recognized Japane … More se Kanji characters of over 90% of the sample cards. A knowledge based and integrated system, which can recognize characters and categorize each bibliographic content field from digitized image of a Japanese language catalog card, is prototyped and tested. The knowledge is stated in the form of frame model. Although the knowledge base has to be refined and tested further, the system has potential of wide applicability to any formatted text printed on paper. 2) An experimental online interactive cataloging simulator program replying a message and uploading local data in place of live cataloger, was implemented on a cataloging workstation for the NACSIS online system. This is used for the evaluation of strategies of automated retrospective cataloging for the union catalog database. Uploading data were obtained through searching and downloading J-Bisc(CD-ROM file). Tests for automatic searching and identifying a bibliographic record in the NACSIS database were carried out. 3) An integrated multimedia database system which combines digitized image file of original documents or original script catalog cards with indexed file was implemented on a conventional optical disk storage and filing system. Less