内容简介
Thevisualrecognitionproblemiscentraltocomputervisionresearch.Fromroboticstoinformationretrieval,manydesiredapplicationsdemandtheabilitytoidentifyandlocalizecategories,places,andobjects.Thistutorialoverviewscomputervisionalgorithmsforvisualobjectrecognitionandimageclassification.Weintroduceprimaryrepresentationsandlearningapproaches,withanemphasisonrecentadvancesinthefield.ThetargetaudienceconsistsofresearchersorstudentsworkinginAI,robotics,orvisionwhowouldliketounderstandwhatmethodsandrepresentationsareavailablefortheseproblems.Thislecturesummarizeswhatisandisn\'tpossibletodoreliablytoday,andoverviewskeyconceptsthatcouldbeemployedinsystemsrequiringvisualcategorization.,