Multimedia Database Management Systems

3y ago
48 Views
2 Downloads
429.08 KB
6 Pages
Last View : 1m ago
Last Download : 2m ago
Upload by : Ellie Forte
Transcription

MultimediaDatabaseARIFGHAFOO17Schoolof Electrical& ComputerManagementEngineering,Purdue1. INTRODUCTIONDue to the heterogeneousnatureof multimediadata, systems intendedto store,transport,display and in general managesuch data must have ation-managementsystems. In this paper,we elaborateon issues pertainingto multimediadatabasemanagementand discuss technicalchallengesinvolvedin developinga general-purposemultimediaDBMS. We can view these issues as specificationrequirementsfor this DBMS andfocus our discussionon a referencearchitecture for such a system. A list of prominent issues in multimediadatabase managementfollows.Developmentof formalsemanticmodeling techniquesfor multimediainformation, especiallyfor video and image data.These models should be rich in capabilities for ics.Theyshould be able to providecanonicalrepresentationsof compleximages,scenes,and events in terms of objects and theirspatio-temporalbehavior.These modelsneed to be comparedand evaluated.Designof powerfulindexing,searching, and organizationmethods for multimediadata.Searchinmultimediadatabasescan be computationallyintensive, especiallyif content-basedretrievalis needed for image and video data storedin compressedor uncompressedform.Developmentof modelsfor rements.Integrationof these modelswiththe monomediadatabaseschemaSystemsUniversity,West Lafayette,Indianawill be required.Subsequently,in orderto determinethesynchronizationrequirementsat retrievaltime,transformationof thesemodelsintoa metaschema is needed. This entailsdesigningobject-retrievalalgorithmsfor the operatingsystems.Similarly,integrationofthese models with higher-levelinformation abstractionssuch as hypermediaorobject-orientedmodels may be hese languagesshould havestrong capabilitiesto express arbitrarilycomplexsemanticandspatiotemporalschemas associatedwith compositemultimediainformation.They must supportmanipulationof content-basedfunctionsfor multimediaobjects.Developmentof efficientdata-placement schemas for physicalstorage management.These schemesare neededtomanagereal-timemultimediadata, forboth single and paralleldisk systems.Design and developmentof suitablearchitectureand operatingsystem support.Heterogeneityof multimediainformationdictatesthat the architectureof a general-purposemultimediadatabasemanagementsystem must supporta rich setof data-managementand talsosupportreal-timerequirements of multimediadata.Managementof distributedmultimediadatabases.In a networkedenvironmentextensivecoordinationand managementcapabilitiesare neededamongthe distributedsites to providelocation-trans-Permissionto make digital/hardcopy of part or all of this work for personalor classroomuse is grantedwithoutfee providedthatcopies are not made or distributedfor profitor commercialadvantage,thecopyrightnotice,the title of the publicationand its date appear,and notice is given that copyingis bypermissionof ACM,Inc. To copy otherwise,to republish,to post on servers,or to redistributeto lists,requiresprior specific permissionand/ora fee.@ 1995 ACM 0360-0300/95/1200-0593 03.50ACMComputingSurveys,Vol. 27, No 4, December1995

594“ArifGhafoorparentaccess and supportlivery of data to distributed2.real-timeusers.A REFERENCE ARCHITECTUREMULTIMEDIADBMSde-FORMany types of real-worldknowledgecanbe representedby describingthe interplay amongobjects (persons,buildings,events,etc.) in the course of timeandtheir relationshipsin space. An application may need to store and access informationabout this knowledgethat can beexpressedbycomplexspatiotemporallogic. A video database is a typical replicaof this real-worldenvironment[Day et al.19951. The s atiotemporalspecificationprocess is rev rsed whhe com osingmultimediainformation/documents[Littleand Ghafoor1993; Iino et al. 1994]. Inthis case, a user syntheticallycreates inter lavamorwvariousmonomediaobjec s iakinto physicalobjects),both inspace and time, In multimediadatabasesthese objectsmay representindividualdata entitiesthat serve as componentsofsome multimediadocument[LittleandGhafoor1993: Stottsand an be groupedtogetherfor efficientmanagementand access.In eithercase, it is essentialthat theuser be able to identifyand address different obiects and be able to ex resstheir.relations”in time and space. These relations shouldbe representablein a suitable structurethat is powerfulenouszh tospecify higher-levelcontentsand semantic abstractions.It is, therefore,desirablethat a generalframeworkfor spatiotemporal aidsemanticmodeling hould beavailable,Such a frameworkcan providea reasonableapproachto address the issues discussedin the .tmevioussectionand can be used in designinga generalpurposemultimediadatabase-management system. Variousconceptualmodelsfor multimediainformationhave beenproposedin the literature[LittleandGhafoor1993; Stottsand Furuta1989].These modelsare eitheraimedat synchronizationaspectsof the multimediadata or are concernedwith the browsingACMComputmgSurveys,Vol. 27, No, 4, December1995aspects of information.Irrespectiveof themedia type, it is imperativethat a modelshould be able to generatea clear specification of the meta-schemaand it must beintegrablewith the underlyingdata models of variousmonomedia,Such a modelmust also facilitatethe efficientdevelopment of high-leveluser interfaces.From this discussion,we can perceivea referencearchitecturefor a generalpurposemultimediaDBMSas shownin Figurel(a). The architectureconsistsof three layers:a monomediadatabasemanagementlayer, a multimediacomposition/managementlayer,and a userinterfacelayer.We now describefunctionalitiesof each layer and discuss theirrole in managingthe issues mentionedinthe previoussection.The monomediadatabase-managementlayer providesthe functionalitiesessential for managingindividualmedia. Thekey aspects of each DBMSat this and allowingusers to de:velopsemantics-basedmodelingandgroupingof complexinformationassociated with each medium.In this respect,this layer deals with the first two issuesdiscussedin the previoussection.Theprimaryobjectiveis to process contentbased queriesas discussedin the previous section.The primaryobjectiveis toprocess content-basedqueriesand facilitateretrievalof appropriatepiecesofmonomediadata.such as videocli s.parts of an image: or audio segments.Forexample,for video data, abstractionhierarchiesin s ace and time.as shown inFigurel(b),’ are needed in order to express events and allow gic can providea formalframeworkforrepresentingsuch eventsand buildingthese hierarchicalabstractions[Day etal. 1995]. Subsequently,this process canlead to schema definitionfor each monomediadatabase.Such schema may alsocontainsome unconventionalinformation, for example,the qualityof service(QoS) parametersfor presentation,suchas speed,volume,resolution,delaybounds,and the like. The objectiveis to

596 ArifGhafoorallow retrievalof monomediadata withthe desiredquality,if possible.Accordingly,a suitabledatabaselanguageisneeded at this level to manipulatehierarchicalabstractionsand to query aboutevents of interest.It is importantto mention that considerablecomputationaldemand can be placed at this layer, especially for imageand video data, if a fullyautomatedsystem is desiredthat is capable of generatingcontent-basedindicesbased on object recognition[Guidavadaand Raghavan1995]. Such computationcan span a broadspectrumof effectiveand robustcomputer-visionand imageprocessingalgorithmsthat are intermingled withdata and entitiesat variouslevels of abstractionsof hinthis layeris the physicalmanagementof individualdatabases.Such managementmay require,as mentionedpreviously,efficientplacementofdata on a single/paralleldisk system.For example,due to the time-variantcharacteristicsof compressedvideo, it isdifficultto predictthe disk seek latenties. A poor data-placementstrategycanresult in significantdegradationin quality due to excessivedroppageof videoframesduringmultipleconcurrentsessions.The primaryobjectiveof the middlelayer is to deal with the thirdissue, theintegrationof monomediato composemultimediadocumentsand to reis a growingdemandfor managementof multimediadocumentsand libraries,and the need forefficientintegrationmodels is one of thekey researchissues in developinga general-purposemultimediaDBMS.Integrationof mediacan span .Therefore,this layer needs tomaintaina multidimensionalmonomediaintegrationmodel,in formof a metaschema, along with the QoS parametersassociatedwith each medium.Basically,the ainformationandgeneratesappropriatesubqueriesfor theACMComputingSurveys,Vol. 27, No. 4, December1995monomediadatabases.Retrievalof themonomediadata is controlledby the integrationmodel maintainedby the metaschema [Littleand Ghafoor1993]. It isimportantto note that for the purpose ofconsistency,this meta-schemaneeds tobe interfacedwiththe schemasassociated with the monomediadatabasespresent at the lowerlayer,becausebothschemasshare informationabout monomedia,includingtheircontent-basedsemanticsand QoS parameters.Anotherimportantfunctionof thislayeris to providecoordinationamongmonomediadatabasesif theyare distributed.As mentionedbefore, one of themajor objectivesin this case is to providelocation-transparentaccess to differentdatabasesitesand maintainsynchronizationamong media streamsoriginating from heterogeneoushosts. From thedatabase point of view, informationaboutthe locationof variousobjects of composite multimediaand their schema must bemaintainedby this layer.The interactivelayer consistsof various user-interfacefacilitiesthat supportmultimediapresentationfunctionalitiessuch as display of images, playout of videoclips or some audiosegment,and thelike. These interactivefacilitiesmay requiresome formalquery languagethatcan be identicalto that used at the bottom layer, with some enhancedcapabilities to vely,some graphical query interfacecan be used for thatpurpose.Additionalcapabilitiesmay include browsing,media editing,and so TECTUREThe developmentof a esign of multilevelmeta-schema,managementof a large numberof indices andphysicaldatabases,as well as interfacingwith the operatingsystem for real-timeretrieval,can be a technologicallydaunting task. In this section, we focus on twodevelopmentissues thatare uniquetomultimediadatabasesand are crucial for

MultimediaDatabaserealizingthe referencearchitectureofFigurel(a): diaandthe suitabilityof two ject-oriented,for developinga generalpurpose multimediaDBMS.A numberof attemptshave been madeto on.HyTime,onesuch model,has been recommendedasan ISO standard.However,currentlythismodelsuffersfromthe drawbackthatthe extractionof variousspatiotemporaland contentsemanticsfrom this modelcan be quite cumbersome.On the otherhand,thePetri-net-basedmodelproposed in Littleand Ghafoor[1993]notonly allows extractionof the desiredsemanticsand generationof a databaseschema in a ratherstraightforwardmanner, but also has the additionaladvantage of pictoriallyillustratingsynchronizationaspectsof the information.Inthis regardthis model is uniqueand isthus also well suited for the third layer ofthe referencearchitecture,where visualorchestrationof multimediainformationcan be highly desirable.The model uses aset of generalizedtemporaloperators.Anexpandedversionof the modelalso allows specificationof spatialsemanticsofinformation[ Iino et al. 1994]. The notionof generalizedspatiotemporalrelationshas recentlybeen extendedto specifycomplexevents in video data, a processthat is almost a reversalof the multimedia compositionprocess [Day et al. 1995].Numerousextensionsto this model havebeen proposed in the literature,includingprovisionfor user interaction,modelingof distributedobject synchronization,developmentof synchronizationprotocolsfor communication,and the like. In summary, the Petri-net-basedmodel in Littleand Ghafoor[1993] can be a good candidatefor spatiotemporalmodelingandgenerationof meta-schemasfor the reference architecture.We now discussthe second developmentissue. As the applicationdomainfor a given multimediaapplicationcan behighlycomplexand ill-structured,bothManagementSystems 597relationalcalculusand semantic-basedobject-orientedapproachesneedto bescrutinized.We cam evaluatethese approaches on two importantcriteria:theirabilityto ions,whichneed to be managedby almostall themonomediaDBMSSof the bottomlayerof Figurel(a), and their expressivepowerfor the specificationof meta-schemaformedia integration,which is managedbythe middlelayer.Hierarchicalabstractions can result in considerablesemanticheterogeneity,which has been a difficultprobemfor relationaldatabasemodels.On the other hand,the er set of abstractionsthat are particularlyuseful for the users to extract\define view of informationat variouslevelsof these hierarchies[Day et al. 1995].From the media integrationand composition point of view, the relationaltechnology againprovidesno elegantmechanismsfor ompositemultimediaobjects. Althoughsome relational DBMSS supportaccess to multimedia objects by using pointersto BLOBS(binarylarge objects),there is no provision for interactivelyaccessingvariousportionof these objects, because a BLOBis treatedas a singleentityin its entirety.It has been demonstratedthat theobject-orientedtechnologycan provideapowerfulparadigmto meet the requirements of multimediacompositionand developingsemanticmodels[Dayet al.1995; Iinoet al. 1994].Its fectivemechanismsfor mediasynchronization.REFERENCESDAY, Y. F., DAGTAS, S., IINO, M., KHOKHAR, A., AND1995.Object-orientedconcepGHAFOOR, A.tual modelingof video data. In Proceedingsofthe EleventhIEEE iwan)March,401-408.GUIDAVADA, V. N. AND RAGHAVAN, V. V., EDS.1995.Content-basedimage retrievalsystems. Specialissue. IEEEComput.28, 9 (Sept).ACMComputingSurveys,Vol27, No. 4, December1995

598Arif GhafoorIINO, M., DAY, Y. F., AND GHAFOOR, A.1994.Anobject-orientedmodel for spatio-temporalsynchronizationof multimediainformation.InProceedingsof the FirstIEEEInternationalConferenceon MultimediaComputingand Systerns (Boston,MA) May, OOR,A.1993.Vol. 27, No 4, December1995Interval-basedconceptualmodelsfor time-dependentmultimediadata. IEEETrans. Knowl.Data Eng, 5, 4 (Aug), 551-563.STOTTS, P. D. AND cturewithbrowsingsemantics.ACMTrans.Oftlce Autom. Syst. 7, 1 (Jan.), 3-29.

multimedia documents and to crosslink information stored in monomedia databases. There is a growing demand for management of multimedia docu-ments and libraries, and the need for efficient integration models is one of the key research issues in developing a gen-eral-purpose multimedia DBMS. Inte-gration of media can span multiple di-

Related Documents:

Introduction to Multimedia (continued) Multimedia becomes interactive multimedia when a user is given the option of controlling the elements. Interactive multimedia is called hypermedia when a user is provided a structure of linked elements for navigation. Multimedia developers develop multimedia projects.

Learn the phases involved in multimedia planning, design and production; Be able to use various multimedia authoring tools Be able to design and create interactive multimedia products Develop competencies in designing and producing instruction-al multimedia Apply contemporary theories of multimedia learning to the development of multimedia .

MULTIMEDIA TECHNOLOGY UNIT – I Multimedia an overview: Introduction The word ‗multimedia‘ comes from the Latin words multus which means ‗numerous‘ and media which means ‗middle‘ or center. Multimedia therefore means ‗multiple intermediaries‘ or ‗multiple means‘. Multimedia

MULTIMEDIA V.S MULTIMEDIA INTERAKTIF Multimedia adalah penggunaan berbagai jenis media (teks, suara,grafik,animasi,danvideo). Multimedia interaktif menambahkan elemen ke-enam yaitu aspek interaktif Pada multimedia non-interaktif, user bertindak pasif dan menyaksikan adegan demi adegan secara berurut

multimedia contexts and for converting one file format to another. Multimedia Editing Tools- These tools are used for creating and editing digital multimedia data. Multimedia Authoring Tools- These tools are used for combing different kinds of media formats and deliver them as multimedia contents. Graphic and Image Editing Software

Multimedia database management systems manage multimedia data including text, images, audio and video. More and more multimedia data are now available on the web and effective management of this data is becoming a critical need. We also need to ensure that the data is protected from unauthorized access as well as malicious corruption.

Database Applications and SQL 12 The DBMS 15 The Database 16 Personal Versus Enterprise-Class Database Systems 18 What Is Microsoft Access? 18 What Is an Enterprise-Class Database System? 19 Database Design 21 Database Design from Existing Data 21 Database Design for New Systems Development 23 Database Redesign 23

Multimedia Systems; Buford; Pearson 3. Multimedia: Sound and Video by Jose Lozano, PHI 4. Multimedia Systems,Tech. & Communications; S.Pandey, M.Pandey; Katson. Chapter 1 Applications of Computer Graphics & Multimedia Copy Right DTE&T,Odisha Page 1 1. Applications of Computer Graphics & Multimedia