Páginas:114Géneros:12:UY:Computerscience12:UT:Computernetworking&communicationsSinopsis:Visualinformationretrieval(VIR)isanactiveandvibrantresearcharea,whichattemptsatprovidingmeansfororganizing,indexing,annotating,andretrievingvisualinformation(imagesandvideos)fromlarge,unstructuredrepositories.ThegoalofVIRistoretrievematchesrankedbytheirrelevancetoagivenquery,whichisoftenexpressedasanexampleimageand/oraseriesofkeywords.Duringitsearlyyears(1995-2000),theresearcheffortsweredominatedbycontent-basedapproachescontributedprimarilybytheimageandvideoprocessingcommunity.Duringthepastdecade,itwaswidelyrecognizedthatthechallengesimposedbythelackofcoincidencebetweenanimagesvisualcontentsanditssemanticinterpretation,alsoknownassemanticgap,requiredacleveruseoftextualmetadata(inadditiontoinformationextractedfromtheimagespixelcontents)tomakeimageandvideoretrievalsolutionsefficientandeffective.Theneedtobridge(oratleastnarrow)thesemanticgaphasbeenoneofthedrivingforcesbehindcurrentVIRresearch.Additionally,otherrelatedresearchproblemsandmarketopportunitieshavestartedtoemerge,offeringabroadrangeofexcitingproblemsforcomputerscientistsandengineerstoworkon.Inthisintroductorybook,wefocusonasubsetofVIRproblemswherethemediaconsistsofimages,andtheindexingandretrievalmethodsarebasedonthepixelcontentsofthoseimages--anapproachknownascontent-basedimageretrieval(CBIR).Wepresentanimplementation-orientedoverviewofCBIRconcepts,techniques,algorithms,andfiguresofmerit.MostchaptersaresupportedbyexampleswritteninJava,usingLucene(anopen-sourceJava-basedindexingandsearchimplementation)andLIRE(LuceneImageREtrieval),anopen-sourceJava-basedlibraryforCBIR._,
"Sinopsis" puede pertenecer a otra edición de este libro.