屬性論指導(dǎo)、含有反饋機(jī)制、基于內(nèi)容的圖像分類

1、上海海事大學(xué)碩士學(xué)位論文屬性論指導(dǎo)、含有反饋機(jī)制、基于內(nèi)容的圖像分類姓名：徐凱申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：計(jì)算機(jī)軟件與理論指導(dǎo)教師：馮嘉禮20070701ABSTRACTWt脅theoverarchingdevelopmentofInternetretievatindustrytheseyears，limitationofattentiononwebsiteretievalhasreadybeentransendedintermsofret

2、rievalcontentFletievalsofsuchcomplicatedimformationasmusicimage。vedioareconstantlypoppingup，whichleadstotheextensiveapplicationofimageretievaltechniqueAtreasentsincethethetextbasedpicturererievaltechnique，whichreliesOnke

3、ywordasacluewithlessresourcespending，isadoptedinallthemainstreamsearchengines。contentbasedretievalpicturetechnique、『I，ithmorepreferablesearchresultsismoreadaptivetothedemandofintemetimageretievaldevelopmentandhaspratlcal

4、significenceThispaperrealizesthecosteffectivepseudosemanticcontentbasedimageretrievalsystem，whichiscalledIntemetorientedPictureRetrievalTechnique(IP明inthispaperAttributetheoryproposedbyProfFeng，isamethodofartificialintel

5、ligenceundertheguidelineofruleofqualititiveandquantitivechangeindialecticmaterialism骶theoryaimsatfindingouttheessenceofobjectviaitsqual黼veandquantitiveattributeandputforwardaseriesofmethodsofpattemrecognition。machineleam

6、ing，evaluationanddecisionmakingectonthebasisofqualititivemappingThispaperUSeSattributetheoryasreferencewhenbuildingfeatureextractionalgorithmofIPRTsystemandappliesthe。analysisapproachofattributecoordinate。intotheretrieva

7、lofuserspsychologicalpreferenceThemainaspectsdealtwffhthispaperarestructuredasfollows：，Proposingtheconceptof。pseudosemanticretdval‘DesigningandrealizingthecompletedIPRTsystem，akindofCBIRsystemfocusingitsretrievalontheint

8、egrationoffourfeaturesascoiourtexture。shapeandsketchwithpreferableaccuracy，highercoverageandfasterexecutionspeed2Improvingtheimagefeatureextractionalgorithm。Adop耄ng搬eHSVcolourspace。whichjsmorefittedtoocularimage。toexpres