280 likes | 454 Views
Making metadata for Manga and applying Topic Maps to it. Shoichiro HARA Center for Integrated Area Studies, Kyoto University Motomu NAITO Knowledge Synergy Inc . Outline. Background and Purpose Functions of Manga Metadata Components of Manga Metadata Apply Topic Maps to the metadata
E N D
Making metadata for Manga and applying Topic Maps to it ShoichiroHARA Center for Integrated Area Studies, Kyoto University MotomuNAITO Knowledge Synergy Inc.
Outline • Background and Purpose • Functions of Manga Metadata • Components of Manga Metadata • Apply Topic Maps to the metadata - Bibliographical Metadata - Structural Metadata 5. WebApplications on topic maps 6. Conclusion 7. Challenge and Future work
1. Background and Purposes • Manga has already established itself a genre of culture • Sales in 2011: 271.7 Bill. / 1112.3 Bill. • One of foremost Japanese pop culture • World wide market • Attractive to Japanese culture and language • Media characteristics • Media diversity/derivation: Weekly periodicals, books, animations, movies, dramas, theaters, parodies etc. • Problems of Manga • Easy to be disappeared • Easy to be out of print and/or cease publication • Easy to be discarded • Easy to be destroyed • Importance of Metadata, digitization and preservation • But there are no well established metadata for Manga database • THEN, our trial to define new metadata of Manga
2. Functions of Manga Metadata • Ability to describe media diversity and derivation • From weekly periodicals to books and animations etc. • Ability to keep all aspects of Manga • Bibliographic, texts, images, movies etc. • Importance of participatory data creation • No special librarians/catalogers of Manga • Need grass roots supports (expressly for foreign materials) • Importance of effective use of Web information • Much information about bibliographies, explanatory notes, authors, publishers etc. are available from Web • Metadata will be used to organize existing information Thus Our Metadata • should be easy to create and use • should be compatible with other metadata
3. Components of Manga Metadata • Bibliographic Metadata • DC based simple structure • Describes elementary information about general concept of work, each work/media (such as book, novel, TV drama, movie, etc.), each resource, etc. • Structural Metadata • Describes structural component of Manga book such as Page, Coma(Frame), Picture, Text, Symbol, etc. • Will try to apply TEI • Topic Maps and Web application • Express metadata as network • Link each component with other Web data
Materials • Title:花より男子 (Hana yoridango) / Boys over Flowers • Author:神尾葉子/ Kamio, Yoko • Outlines • Original: Serial Manga published inマーガレット (Marguerite), a biweekly girls’ Manga magazine published by 集英社 (Shuei-sha), from 1992 to 2004. • Book: 37 volumes. The best selling girls’ Manga in Japan (58000000 copies until Sep. 2006) • Multimedia: animation, drama, movie. • Translated versions: USA, Taiwan, Thailand, France, Spain etc. • Reasons for Research Material • Various medias: good sample for metadata • Various translated versions: good for multilingual text sample • Material : Book version (Vol. 1) • Japanese (original) (Shuei-sha) • English (Viz Media) • Thai (Siam Inter Comics)
Image Data • Image scanning • 300dpi • Gray scale(8bit), Only cover page is color scanning • Preprocessing • Remove the cover • Then cut the binding of the book (Bind the book again after scanning)
4. Apply Topic Maps to the Metadata What and Why Topic Maps? • Simple, Intuitive and Human-friendly model for organizing information • Models target domain as concept/subject network (topics and associations between topics) • Links concepts and related information resources (occurrences) • Consists of types (which corresponding Ontology) and instances • Same vocabularies and syntax are used for Types and Instances • Uses Subject Identifier (IRI or URI) to identify subjects (topics) and link/merge subjects • Has very powerful Topic Maps query language: tolog • Has remote access protocol: TMRAP to access remote topic maps • Topic Maps is an ISO standard (ISO/IEC 13250) • Can use matured open source tools, e.g. Ontopia http://code.google.com/p/ontopia/
Some pool of information or data • any type, format, or location draw • A knowledge layer, consisting of: draw • Topics • a set of topics representing the key subjects of the domain in question Cat Street BOYS over FLOWERS • Associations • representing relationships between subjects Yoko Kamio born in TOKYO • Occurrences • links to information that is somehow relevant to a given subject knowledge information The Basic Model of Topic Maps P.S.Topics, associations and occurrences have types, and all types are also topics… • = The TAO of Topic Maps (Source: Steve Pepper, “Towards Seamless Knowledge” )
General View and target of Manga Metadata • Three types targets • Work(著作), Expression(表現形), Manifestation(体現形) • Metadata is made for each instance of the types • Describing relationships (e.g. derivation) between instances
Bibliographic Metadata • Metadata is described by DCMES’s(Dublin Core Metadata Element Set) 15 elements and relationship between them • Attaching identifier(IRI or URI) to the elements as far as possible Refer to: Steve Pepper, Expressing Dublin Core Metadata using Topic Maps, http://www.jtc1sc34.org/repository/0906.htm
Topic map of Bibliographic Metadata • Topic Types(8): Work, Expression, Manifestation, Agent, Coverage class, Language class, Subject class, Type class • Occurrence Types (5): Description, Date, Format, Identifier, Rights • Association Types (9): createdBy, hasSubject, publishedBy, contributedBy, typeAs, sourceOf, expressedIn, relatedWith, coveredWith • Association role Types (18): creator, subject, publisher, contributor, type, source, language, relation, coverage, resource-role-for-creator, resource-role-for-subject, resource-role-for-publisher, resource-role-for-contributor, resource-role-for-type, resource-role-for-type, resource-role-for-language, resource-role-for-relation, resource-role-for-coverage • title is used as topic name of resources (Work, Expression, Manifestation) * Red characters represent DCMES
Structural Metadata • Structural Metadata is described by Composition elements of Manga (volume, scene, page coma, script/onomatopoeia, person) and relationship between them • IRIs (URIs) are attached to all components
Topic map of Structural Metadata • Topic Types(6):Volume, Scene, Page, Coma, Script/Onomatopoeia, Person • Occurrence Types (3):Attribute, String, image • Association Types (8):part-of, appear, depict, speak, before-after, coma-before-after, script-before-after, scene-before-after • Association Role Types (16):whole, part, coma-appearing-person, person-in-coma, coma-depicting-script, script-in-coma, speaker, script-of-speaker, before, after, coma-before, coma-after, script-before, script-after, scene-before, scene-after • Scope : Japanese(ja), English(en), Thai(th)
Topic map of Structural Metadata • Instance topics • volume: 1 • scene: 29 • page: 158 • come: 550 • script/onomatopoeia: 1,239 • person (character): 18 • Instance associations • part-of: 780 • appear: 520 • depict: 1239 • before-after: 157 • coma-before-after: 547 • Script-before-after:1238
5. WebApplications on topic maps • Make Bibliographic topic map and Structural topic map separately • Transform CSV format to Topic Maps syntax • Using “DB2TM” in Ontopia as a tool • Make Web application separately • Using “Ontopia Navigator Framework” in Ontopia as a tool • Link bibliographic topic map with structural topic map • One way link at the moment
Functions of the Webapplications • Instance list for each topic type • Instance details • Topic characteristics(topic name, internal occurrence, external occurrence, association role) • Graphic representation • Keyword search • Tolog query • Navigating topics based on associations • Web service interface by TMRAP
Web Application for Structural Metadata Top page Key word search
Structural Metadata Web application Script/Onomatopoeia detail Script/Onomatopoeia list Coma detail
6. Conclusion • Manga and related resources should be preserved and inherited • Metadata plays very important roles to realize it • By our method metadata changes from mere list of elements to structured network of concepts (including derivation) • Necessity for introduction of semantic technology • This research uses Topic Maps technology • Computer can identify concepts using IRI or URI • Queries can be done from various point of views and based on association In result • It becomes easy to manage, find and access the metadata and information resources (target of metadata) • It makes new way to create and add new value to domains
7. Challenge and Future work • Increase examples of Manga Metadata • Conform the Bibliographic Metadata to FRBR • Develop environment for metadata input and maintenance • Make metadata guideline for animation, drama, movie, etc. • Develop environment for register, manage and use of Subject Identifier(IRI or URI) • Use authority data and ID of author and organization • E.g. Web NDL Authorities Yoko Kamio: http://id.ndl.go.jp/auth/ndlna/00320527 • Ensure enough terms for subject (NDLSH are not enough) • Make multilingual metadata • And all that
Challenge and Future work (cont.) Utilization of the Metadata • Send messages about Manga related information resources to all over the world and pass them down the generations • Try to apply the metadata to the learning materials, user guides and maintenance manuals for the instruments, etc. In order to increase usefulness of the Metadata, we would like to make sure some ambiguity requests can also be fulfilled. Such as: • I'd like to find “that manga” which I read when I was a child. • I'd like to use “appropriate manga” to check “Trend, Culture, Fashion, etc.” of “some old days”.
Thank you! Any suggestion?