Real-Time Accumulative Computation Motion Detectors 英文参考文献.docVIP

Sensors2009,9,10044-10065;doi:10.3390/OPENACCESS sensors ISSN1424-8220 /journal/sensors Article Real-TimeAccumulativeComputationMotionDetectors AntonioFern′andez-Caballero 1,2,?,Mar′?aTeresaL′opez 1,3,Jos′eCarlosCastillo 1 and SaturninoMaldonado-Basc′on 4 1 2 3 4 InstitutodeInvestigaci′onenInform′aticadeAlbacete,02071-Albacete,Spain; E-Mails:mlopez@dsi.uclm.es(M.T.L.);josecarlos@dsi.uclm.es(J.C.C.) DepartamentodeSistemasInform′aticos,EscueladeIngenierosIndustrialesdeAlbacete,Universidad deCastilla-LaMancha,02071-Albacete,Spain DepartamentodeSistemasInform′aticos,EscuelaSuperiordeIngenier′?aInform′atica,Universidadde Castilla-LaMancha,02071-Albacete,Spain DepartmentofSignalTheoryandCommunications,EscuelaPolit′ecnicaSuperior,Universidadde Alcal′a,28871-Alcal′adeHenares,Madrid,Spain;E-Mail:saturnino.maldonado@uah.es ? Authortowhomcorrespondenceshouldbeaddressed;E-Mail:caballer@dsi.uclm.es. Received:28October2009;inrevisedform:24November2009/Accepted:30November2009/ Published:10December2009 Abstract: Theneurallyinspiredaccumulativecomputation(AC)methodanditsapplication to motion detection have been introduced in the past years. This paper revisits the fact that many researchers have explored the relationship between neural networks and ?nite statemachines. Indeed,?nitestatemachinesconstitutethebestcharacterizedcomputational model,whereasarti?cialneuralnetworkshavebecomeaverysuccessfultoolformodelingand problemsolving. Thearticleshowshowtoreachreal-timeperformanceafterusingamodel describedasa?nitestatemachine. Thispaperintroducestwostepstowardsthatdirection: (a)Asimpli?cationofthegeneralACmethodisperformedbyformallytransformingitinto a ?nite state machine. (b) A hardware implementation in FPGA of such a designed AC module,aswellasan8-ACmotiondetector,providingpromisingperformanceresults. We alsooffertwocasestudiesoftheuseofACmotiondetectorsinsurveillanceapplications, namelyinfrared-basedpeoplesegmentationandcolor-basedpeopletracking,respectively.