DESIGNOFA VISIONSENSORUSINGFUZZYASSOCIATIVE DATABASE XiangChent*, Shahed Shahirt, andMajidAhmadi
Universityof Windsor, Universityof Waterloo · 2005
Inthispaperadesign methodisproposed fora potential newcamera-based intelligent vision sensor.Thissensor canbeusedforfast multiple planarobject recognition. Themechanism behind thedesign isaFuzzyAssociative Database (FAD) whichconsists ofaFuzzyDatabase (FD)anda FuzzySearch Engine (FSE). TheFSEusestable onetoconduct search overtable two,bothin FD,through aBankofFuzzyAssociative MemoryMIatrix(BFAMM). Infact, theFSEestablishes acorrespondence between anobject andoneof thetrained classes intable two.Therefore, the FAD couldactually 'remember' thetrained ob-_ jects andtheFSEcould 'recognize' theincoming object bycomparing itwithtrained information inthedatabase. Theexperimental results show thatthis approach isrobust andfast.